Lineages of human T-cell clones, including T helper 17/T helper 1 cells, isolated at different stages of anti–factor VIII immune responses

Anti–factor VIII (FVIII) inhibitory antibodies (referred to clinically as inhibitors) develop in 20% to 35% of severe and 3% to 13% of mild/moderate severity hemophilia A patients who receive FVIII replacement therapy, and they are associated with significant mortality, morbidity, and lower quality of life.^1-3  FVIII-specific antibodies are dependent on CD4⁺ T helper (T_H) cells,⁴  which respond to FVIII-derived peptides bound to major histocompatibility complex class II receptors on antigen-presenting cells when these class II-peptide complexes bind with sufficient avidity to their T-cell receptors, causing proliferation and cytokine release. CD4⁺ T cells may be divided into 4 subsets that secrete distinct cytokines, as follows: T_H1, T_H2, inducible regulatory T cells, and the recently defined T_H17 cell lineage.^5,6  These subsets have distinct functions in up- and down-regulating immune responses and in orchestrating antibody class switching.^7-10  CD4⁺ T cells from hemophilia A subjects with inhibitors have been found to secrete interferon (IFN)–γ, indicating the presence of T_H1-polarized cells, and interleukin (IL)–4, indicating the presence of T_H2-polarized cells, upon FVIII stimulation.¹¹  In contrast, FVIII-stimulated T cells from some hemophilia A subjects without inhibitors secreted both the proinflammatory cytokine IFN-γ and the anti-inflammatory cytokine transforming growth factor (TGF)–β, with the latter indicating the induction of inducible regulatory T cells.¹¹  T_H17-polarized cells are associated with chronic inflammation and with various autoimmune disorders,^12,13  but there has as yet been no report of their possible role in hemophilia A immune responses. T-cell clones that respond to distinct epitopes in the FVIII A2, C1, and C2 domains have been isolated from several hemophilic subjects,^14-17  and these monoclonal CD4⁺ cell lines are proving useful in mechanistic investigations of inhibitor development.

T-cell clones isolated from 2 brothers with mild hemophilia A due to the missense substitution A2201P, including clones from serial samples obtained after inhibitor development, showed highly similar HLA-DRA-DRB1*0101-restricted T-cell responses to an epitope in FVIII that included the missense substitution site.^15,16  The objective of the present study was to investigate the phenotypes of these clones by characterizing cytokine production, chemokine receptor expression, and transcription factor usage, as phenotypic changes over time and differences between these subjects should offer clues as to why only one of them developed a clinically significant inhibitor. Thirteen T-cell clones were isolated from the proband's blood at 2 time points, and 6 clones were obtained from a single sample obtained from his brother. Interestingly, T_H17/T_H1-polarized clones were isolated from the sample obtained 19 weeks after initial detection of the proband's inhibitor, whereas clones isolated at the 21-month time point all had a distinct T_H2 profile, suggesting that the T_H17 cell lineage played a role only in earlier stages of this person's anti-FVIII immune response. All 6 clones isolated from the noninhibitor subject were T_H1 polarized, a phenotype previously seen in hemophilic subjects with and without an inhibitor, as well as in healthy, nonhemophilic controls.^11,14,17 

Blood samples from hemophilic brothers with the FVIII missense substitution A2201P, who shared the HLA-DRA-DRB1*0101 allele, were obtained after written, informed consent, according to a protocol approved by the University of Washington Human Subjects Review Committee and the Declaration of Helsinki. In a previous study,¹⁶  the proband and his (half-)brother were designated IV-1 and IV-2, reflecting their assigned designations in the family pedigree. They are referred to in this study as subjects 17A (the proband) and 32A (his brother), respectively, because these clinical study identifiers were also used to label T-cell clones isolated from their blood. The naming convention for the clones consists of the subject number, the amount of time elapsed since initial inhibitor development (if applicable), and the sequential number of the clone isolated from a particular blood sample. The clinical status of the proband and his brother was described previously.^15,16  The proband did not receive any additional FVIII infusions during the course of the present study. His inhibitor titer at 19 weeks was 29 BU/mL, and at 21 months it had fallen to 5 BU/mL.

Isolation of 5 T-cell clones from subject 17A 19 weeks after initial detection of his inhibitor, and of 6 clones from subject 32A, was previously described.^15,16  Eight additional T-cell clones that recognized the synthetic peptide FVIII_2194-2213 were isolated from subject 17A 21 months after initial inhibitor detection using the same method. Clones were expanded by stimulation with irradiated peripheral blood mononuclear cells (PBMCs) from an HLA-mismatched person plus phytohemagglutinin (REMEL) in the presence of human IL-2 (Hemagen Diagnostics; a detailed protocol is given in supplemental methods; available on the Blood website; see the Supplemental Materials link at the top of the online article).

T-cell proliferation assays were performed, as described.¹⁶  Supernatants were collected 48 hours after peptide stimulation and were stored at −80°C until enzyme-linked immunosorbent assays (ELISAs) were performed. Supernatants from triplicate wells were pooled and analyzed by ELISA to measure IFN-γ, IL-4, IL-17A, and TGF-β₁. Sandwich ELISAs were performed using standard methods (supplemental methods).

Total RNA was isolated from 2 million resting T cells using an RNeasy Mini Kit (QIAGEN). RNA samples were treated with DNase using a TURBO DNA-free kit (Applied Biosystems). cDNA was generated using a high-capacity cDNA reverse-transcription kit (Applied Biosystems). Transcripts were quantified on a 7900HT Fast Real-Time PCR System using predesigned TaqMan gene expression assays, according to the manufacturer's instructions (Applied Biosystems). Probes with the following Applied Biosystems assay identification numbers were used: TBX21, Hs00203436_m1; GATA3, Hs00231122_m1; RORC, Hs01076112_ml. mRNA levels were normalized to 18S rRNA levels, as quantified with the 18S probe/primer TaqMan assay kit (Applied Biosystems). Mean cycle threshold values (n = 3) were determined from amplification plots using Sequence Detection System Software Version 2.3 (Applied Biosystems). Relative mRNA expression levels were determined from a standard curve of serial RNA dilutions from clone 32A-14 (TBX21 mRNA, GATA3 mRNA, and 18S rRNA standard curves) and 17A-19WK-11 (RORC mRNA standard curve).

Resting T-cell clones were stained with antibodies, according to the manufacturer's instructions, and flow cytometry was performed on a FACSCalibur (BD Biosciences). Antibodies were from R&D Systems: monoclonal anti–human CXC chemokine receptor (CXCR)3-carboxyfluorescein (CFS) IgG₁, mouse IgG₁ isotype control-CFS, monoclonal anti–human C-C chemokine receptor (CCR)4-CFS IgG_2B, mouse IgG_2B isotype control CFS, monoclonal anti–human CCR6 allophycocyanin, and mouse IgG_2B isotype control allophycocyanin. Mean fluorescence intensities (MFI) were determined using FlowJo software.

Cytokine secretion and proliferation by T-cell clones isolated from the proband, subject 17A, 19 weeks and 21 months after inhibitor detection, and by clones obtained from his brother, subject 32A, 2 years after his last FVIII infusion, were evaluated. Isolation, tetramer staining, and proliferation assays of some of these clones were described previously.^15,16  A total of 19 clones was stimulated with 0.1, 1.0, and 10 μM FVIII_2194-2213: 5 from subject 17A at the 19-week time point, 8 from subject 17A at the 21-month time point, and 6 from subject 32A. All clones showed dose-dependent proliferation in response to peptide FVIII_2194-2213 (Figure 1A). IFN-γ, IL-4, and IL-17 secretion levels identified 4 distinct cytokine secretion profiles (Figure 1B-D). Two profiles were seen for the 19-week 17A clones, a third profile was seen for the 21-month 17A clones, and a fourth profile was seen for all 32A clones. Clones 17A-19WK-5 and 17A-19WK-10 secreted IL-4 and low amounts of IFN-γ (< 300 pg/mL at 10 μM peptide). Clones 17A-19WK-11, -22, and -33 showed dose-dependent secretion of IL-17 and IFN-γ, although the IFN-γ levels were low. All 8 of the 21-month 17A clones secreted predominantly IL-4, and all 6 clones from 32A secreted high levels of IFN-γ and IL-4. T-cell clones cultured with no peptide or with the irrelevant peptide FVIII_519-538 all showed low cytokine production (< 145 pg/mL IFN-γ, < 35 pg/mL IL-4, < 9 pg/mL IL-17; data not shown). TGF-β₁ was not detected in any of the supernatants (data not shown). To verify that IL-17 secretion was not an artifact resulting from culturing cells under conditions that could influence their polarization, IL-17 secretion from clones 17A-19WK-11, 17A-19WK-22, and 17A-19WK-33 was assayed using cells that had been frozen after early passages of the original clonal culture (Figure 2). The cytokine profile was similar to that seen in Figure 1, although somewhat higher levels of both IFN-γ and IL-17 were detected.

IFN-γ, IL-4, and IL-17 secretion levels are shown as ratios to reflect T-cell polarization relative to “pure” T_H1, T_H2, or T_H17 cytokine responses (Figure 1E-F). Cytokine ratios were calculated after stimulation with 0.1, 1.0, and 10 μM FVIII_2194-2213. Clones 17A-19WK-5 and 17A-19WK-10 were either T_H1/T_H2 polarized or T_H2 polarized, and clones 17A-19WK-11, 17A-19WK-22, and 17A-19WK-33 were T_H17/T_H1-polarized, whereas at 21 months after initial inhibitor development only T_H2-polarized clones were isolated. Calculated ratios were variable for 17A-21MO clones, as IFN-γ levels were at or slightly above background levels. The clones from the tolerant subject 32A secreted approximately 10 times as much IFN-γ as IL-4, consistent with T_H1 polarization, although dose-dependent IL-4 secretion was still significant.

Cytokine secretion patterns indicated the T-cell clones had distinct T_H lineages. To better characterize their phenotypes, mRNA expression levels of the lineage-specific transcription factors TBX21, GATA3, and RORγt, which are used by T_H1, T_H2, and T_H17 cells, respectively,^5,6  were measured using TaqMan gene expression assays (Figure 3A). The RORγt mRNA assay measured both isoforms expressed from the RORC gene. Cell surface expression of chemokine receptors CXCR3, CCR4, and CCR6, which are associated with T_H1, T_H2, and T_H17 effector function, respectively,^18-20  was also measured by flow cytometry (Figure 3B). Clones 17A-19WK-5 and 17A-19WK-10 expressed TBX21 and GATA3 mRNAs, as well as CXCR3 and CCR4 proteins, consistent with the T_H1/T_H2 lineage deduced from their cytokine profiles. Clones 17A-19WK-11, -22, and -33 expressed RORC and TBX21 mRNA, as well as high levels (MFI ∼100) of CCR6 protein and moderate levels of CXCR3 and CCR4, consistent with a T_H17/T_H1 lineage. The 17A-21MO clones expressed high levels of GATA3, low levels of TBX21, low levels of CCR4 protein, and CXCR3 expression just above background; despite the relatively low CCR4 expression, this profile was consistent with a T_H2 lineage. The 32A clones exhibited some heterogeneity, with various amounts of TBX21 and GATA3 mRNA. Expression of CXCR3 was lower on clones 32A-14 and 32A-21, consistent with their lower level expression of TBX21. Three of these clones also expressed low to moderate levels of CCR6, which was observed on only a fraction of the cells. Histogram plots showed 2 distinct populations for these clones, only one of which expressed CCR6 (data not shown), even though more than 95% of these cells were stained by DR0101 tetramers loaded with peptide FVIII_2194-2213. This indicated that the 2 distinct populations revealed by CCR6 staining did not lose clonality during expansion. Overall, phenotypes for clones 32A-5, 32A-15, 32A-16, and 32A-18 indicated a T_H1 polarization, whereas clones 32A-14 and 32A-21 showed a profile indicating a mixed T_H1/T_H2 lineage.

The cloning of T cells restricted by the same HLA-DR allele from 2 brothers with mild hemophilia A, only one of whom developed a high-titer inhibitor, presented an opportunity to explore characteristics of human T-cell responses to a clinically relevant FVIII epitope. T cells from both subjects responded to one epitope that coincides with the A2201P missense substitution site.¹⁶  Distinct patterns of cytokine secretion, transcription factor mRNA, and chemokine receptor expression were seen for clones isolated from the proband at 2 time points, approximately 5 and 21 months after he developed an inhibitor. Clones isolated from his brother, who did not have a clinically significant inhibitor, showed yet another profile. Our results indicated that the T_H lineages of FVIII-responsive cells changed over time after an initial inhibitor response, and a separate T_H lineage was found for clones isolated from the subject who has maintained B-cell tolerance to FVIII.

To our knowledge, the clones described in this study are the first evidence that FVIII-specific T_H17/T_H1 cells are present in an inhibitor response. The detection of T_H17/T_H1 cells at 19 weeks, but not 21 months, suggests that T_H17/T_H1 lineage cells play a role in the early months of anti-FVIII antibody production, but that they were not involved in maintenance of long-term B-cell responses to FVIII. Interestingly, a study of T-cell responses in 2 unrelated subjects with mild hemophilia A due to missense substitution R593C, both of whom had a longstanding low-titer inhibitor after an earlier peak response, characterized FVIII-specific T-cell clones, and they did not secrete IL-17.¹⁷  Because these blood samples were obtained years after initial inhibitor development, this result is consistent with the notion that T_H17/T_H1-polarized cells may play a role in earlier stages of the immune response. Additional studies analyzing responses of T cells isolated from serial samples after an inhibitor response are needed to verify and properly interpret the discovery of this proinflammatory T-cell lineage in the present study.

T_H17 cells play an important role in host defense against infection and in tissue inflammation by inducing the release of proinflammatory and neutrophil-mobilizing cytokines.^12,13,21  The proband developed an inhibitor after receiving FVIII to support tonsillectomy/adenectomy, so T_H17 cells having specificities for other antigens were most likely already present when he received the initial infusions with wild-type FVIII. IL-21 produced by T_H17 cells is a powerful positive-feedback stimulant of T_H17 cell differentiation.^5,6,13  Thus, the cytokine environment may have promoted differentiation of naive T cells toward a T_H17 polarization. Inflammation is a widely acknowledged risk factor for inhibitor development; therefore, it is reasonable to hypothesize that the T_H17-polarized cells may contribute to the development of anti-FVIII antibodies.

T_H17 and T_H17/T_H1 cells, once formed, have the capacity to stimulate antibody production and regulate humoral immune responses.^22-27  In various mouse models of allergy and autoimmunity, IL-17 deficiency decreases germinal center B-cell development and antigen-specific antibody production, whereas high levels of IL-17 promote germinal center and antigen-specific antibody production.^22-24,26,27  Human T_H17 and T_H17/T_H1 clones have been shown to induce CD19⁺ B cells to produce IgM, IgG, and IgA, but not IgE when stimulated with anti-CD3 antibody.²⁵  Human T_H17 cells, but not T_H1 and T_H2 cells, secrete high levels of the B-cell chemoattractant CXC chemokine ligand 13.²⁸ 

This is the first description of antigen-specific, HLA-restricted, T_H17-polarized human clones isolated using culture conditions optimized for T_H1- and T_H2-polarized cells. Takagi et al²⁸  produced human T_H17 alloreactive and Candida albicans-specific clones by culturing CD4⁺ cells under T_H17-polarizing conditions and cloning by limiting dilution. Streeck et al²⁹  isolated T_H17 and T_H17/T_H1 clones by sorting IL-17–secreting cells labeled with an IL-17 capture complex. In the present study, T_H17/T_H1 clones were isolated from CD4⁺ T cells by tetramer staining, followed by single-cell sorting and expansion in the presence of IL-2 and human serum. Their isolation under these conditions was somewhat surprising, because it has been reported that IL-2 inhibits T_H17 polarization of murine T cells.³⁰  Although 2 of the 3 T_H17/T_H1 clones (17A-19WK-22 and 17A-19WK-33) expanded less robustly in the presence of IL-2, expansion of clone 17A-19WK-11 was comparable with that of the T_H2 and T_H1/T_H2 lineages described in this study (data not shown). The T_H17/T_H1 clones, however, had a greater tendency to form aggregates in culture (data not shown). Various culture conditions are now being tested to determine optimal conditions for expansion and maintenance of human T_H17/T_H1 clones.

The identification of FVIII-specific T_H2 and T_H1/T_H2 cells in an inhibitor subject, and of T_H1 cells in a hemophilic subject without an inhibitor, is consistent with an earlier report describing polyclonal CD4⁺ T-cell responses to FVIII by Hu et al.¹¹  In another study, hemophilia A mice that were treated with anti-CD3 to prevent inhibitors showed a shift in antigen-specific cytokine production from a mixed T_H1/T_H2 response toward a T_H1 polarization.³¹  Results of these recent studies and of the present investigation indicate that T_H1 responses to FVIII can persist in hemophilic mice and humans without precipitating development of a high-titer inhibitor.

The T_H lineages of the clones were determined by measuring cytokine secretion in response to antigen, and by measuring transcriptional regulators and surface chemokine receptor expression for resting clones. The mRNA expression levels of the lineage-specific transcriptional regulators TBX21, GATA3, and RORC correlated well with cytokine secretion patterns of the clones. The mRNA expression levels clearly marked the different cell types, particularly in the case of RORC, which was expressed only by clones that secreted IL-17 in response to antigen. The chemokine receptor expression patterns, in contrast, did not strictly correlate with cytokine secretion levels; in particular, CCR4 expression was not appreciably up-regulated on the T_H2 clones and it was most pronounced on the T_H17/T_H1 cells. CCR6 was expressed on the clones that secreted IL-17, as expected, but it was also detected on 3 clones from the noninhibitor subject. Other studies have also noted that chemokine receptor expression is not strictly correlated with other T_H lineage markers. Singh et al²⁰  observed that CCR4 expression was up-regulated on human cord blood CD4⁺ T cells differentiated under T_H17 conditions. Annunziato et al²⁵  examined the expression of CCR4 and CCR6 mRNA in human T-cell clones and found that CCR6 expression was diagnostic of T_H17 and T_H17/T_H1 clones, whereas CCR4 was expressed on cloned cells of various lineages. Acosta-Rodriguez et al³²  showed that CCR6 is expressed on both IL-17 and IFN-γ-secreting human memory T cells from peripheral blood. The majority of these IL-17–secreting cells were CCR6⁺CCR4⁺, whereas IFN-γ–secreting cells were CCR6⁺CXCR3⁺ or CCR6⁻CXCR3⁺. Cells secreting both IL-17 and IFN-γ were either CCR6⁺CCR4⁺ or CCR6⁺CXCR3⁺. Overall, these studies suggest that CCR4 expression is not specific for T_H2 cells, whereas CCR6 is generally expressed by IL-17–secreting cells, but may also be expressed by IFN-γ–secreting cells under certain conditions.

In conclusion, detailed phenotypic characterization of FVIII-specific T-cell clones isolated at earlier and later stages of an inhibitor response, and from a noninhibitor subject who nevertheless had a strong T-cell response to FVIII, indicated changes in T-cell lineages that accompany the progression of immune responses after FVIII stimulation. Additional time-course studies, including prospective studies following the differentiation of naive T cells to various T_H lineages, will be necessary to confirm and extend the findings reported in this study. It will be particularly important to determine whether FVIII-specific T_H17 cells are generally involved in initial stages of helper T-cell responses to FVIII replacement therapy, or whether they contribute to inhibitor development primarily when inflammation is present concurrent with FVIII administration.

We thank Charles Cooper for help with human subjects; Shelley Nakaya for DNA sequencing; Lonna Pells for assistance characterizing T-cell clones; Gerald T. Nepom for critical reading of the manuscript; and all subjects for blood donations.

This work was supported by a Bayer (Berlin) hemophilia award (to K.P.P.), a CSL Behring (King of Prussia, PA) hemophilia research award (to K.P.P.), National Institutes of Health (Bethesda, MD) Grant R01-HL 071093-01 (to A.R.T.), and National Institutes of Health Contract HHSN266200400028C (to W.W.K.).

National Institutes of Health

Contribution: R.A.E. designed and performed the majority of the experiments, analyzed the data, and wrote the manuscript; E.A.J. isolated some T-cell clones, generated tetramers, and edited the manuscript; W.W.K. oversaw tetramer generation and edited the manuscript; A.R.T. oversaw the human subjects and edited the manuscript; and K.P.P. oversaw the study and wrote the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Kathleen P. Pratt, Puget Sound Blood Center, BRI Rm 3016, 921 Terry Ave, Seattle, WA 98104; e-mail: kathleenp@psbcresearch.org.