Background
Acute myeloid leukemia (AML) is characterized by clonal expansion of immature myeloblasts in the bone marrow and eventually leukemization [
1]. The only curative treatment is intensive chemotherapy that can be combined with stem cell transplantation especially in younger patients below 60-65 years of age [
1]. This treatment is followed by a period of severe bone marrow failure with severe pancytopenia, including lymphopenia. Rapid lymphoid reconstitution after therapy is associated with increased AML-free survival, an observation strongly suggesting that immunological events early after chemotherapy are clinically important [
2,
3]. This association between reconstitution and survival has been observed after conventional intensive chemotherapy [
4], autotransplantation [
5] and allotransplantation [
6,
7]. The mechanisms behind these associations are not known but may involve (i) treatment-induced immunogenic cell death with translocation of endo-calreticulin to the cell surface and induction of antileukemic T cell reactivity [
8], (ii) increased efficiency of antileukemic immune reactivity during this period of low leukemia cell burden, or (iii) treatment-induced alterations of immunoregulatory networks. The remaining lymphocytes during chemotherapy-induced cytopenia may thus be important for this antileukemic effect.
Previous studies have described increased levels of circulating immunosuppressive CD4
+ CD25
HIGH T
REG cells for patients with untreated AML [
9,
10] and these increased levels persist after chemotherapy when complete hematological remission is achieved [
9]. The pretreatment T
REG levels even seem to predict the response to the chemotherapy [
9]. The peripheral blood levels of proinflammatory T
H17 cells and IL17 plasma levels are also suggested increased in untreated AML, but in contrast to T
REG cells these levels normalize when complete remission is achieved [
11]. However, the levels of various T cell subsets during the early period of chemotherapy-induced cytopenia have not been investigated previously, and in the present study we therefore compared the T cell subset distribution for AML patients (i) before treatment, (ii) during the period of severe chemotherapy-induced cytopenia, and (iii) during hematopoietic reconstitution after treatment. The T cell subsets investigated were IFNγ secreting CD8
+ cytotoxic T (T
C1) and CD4
+ helper T (T
H1) cells, IL17-A secreting CD4
+ helper (T
H17) cells and CD4
+ CD25
+ FoxP3
+ regulatory T (T
REG) cells.
Discussion
Previous studies of TREG cells in human AML have included patients with untreated disease and patients in complete hematological remission. AML patients with chemotherapy-induced cytopenia have not been examined, even though clinical studies suggest that immunological events early after chemotherapy are important for the antileukemic effect of chemotherapy. In the present study we therefore compared T cell subset distribution in AML patients with severe chemotherapy-induced cytopenia with (i) untreated patients or patients during regeneration after intensive chemotherapy; and (ii) healthy controls. The cytopenic patients had increased relative levels of circulating TREG cells, whereas levels of TC1 and TH1 cells were decreased while TH17 cells were not altered.
Early lymphoid reconstitution after chemotherapy is associated with decreased risk of leukemia relapse [
4‐
7,
28], and these observations suggest that immunological events early after chemotherapy are clinically important. The immunological status during treatment-induced cytopenia are determined by pre-existing disease-induced abnormalities and chemotherapy-induced defects [
21]. Previous studies have demonstrated that even patients with severe therapy-induced lymphopenia have an operative T cell system [
21], and in this context we investigated the balance between various proinflammatory and suppressive T cell subsets in AML patients.
As expected the relative levels of CD3+ T lymphocytes to total leukocytes were dependent on the degree of leukemization in patients with untreated AML. More important, the frequencies of circulating CD3+ T cells among total lymphocytes were decreased in patients with untreated disease compared with healthy controls, these levels were increased during chemotherapy-induced pancytopenia before normalization during reconstitution. Even though our cytopenic patients had lymphopenia, our results suggest that CD3+ T cells are less sensitive to intensive chemotherapy than myeloid cells and other lymphocyte subsets (e.g. B cells, NK cells).
We did not find any significant differences in relative CD8
+ and CD4
+ T cell levels between healthy controls and patients examined before or following intensive chemotherapy. In contrast, the relative levels of circulating IFNγ secreting CD4
+ (T
H1) and CD8
+ (T
C1) T cells were decreased in cytopenic patients before normalization during reconstitution (Figure
3). This decrease is also reflected in the decreased T
C1:T
REG and T
H1:T
REG ratios during cytopenia (Figure
6). Taken together these observations strongly suggest that treatment-induced lymphopenia is not a random process and the susceptibility to intensive chemotherapy differ between T cell subsets.
T
H17 cells constitute a separate proinflammatory T helper (T
H) cell subset [
29‐
33]. Studies in animals as well as humans suggest that T
H17 cells are important for anticancer immune reactivity [
34,
35]. A previous study described increased frequencies of circulating T
H17 cells in patients with untreated AML and with a normalization when patients achieved complete remission after chemotherapy [
11]. In contrast, we observed normal T
H17 levels for patients with untreated AML. In addition we observed that functional T
H17 cells could be detected during severe therapy-induced cytopenia, and these frequencies did not differ from healthy controls although the variation range was broader. Studies of absolute and relative T
H17 levels in a second patient cohort confirmed that only minor variations were seen during therapy-induced cytopenia. The decreased T
H17:T
REG ratio during cytopenia is thus caused by relatively stable T
H17 absolute levels with only minor decreases and maintained or increasing absolute levels of T
REG cells (Figure
6). Finally, we observed increased IL17-A release by circulating cells in response to stimulation by T cell specific antibodies (anti-CD3 combined with anti-CD28); this is probably caused by T cell release because IL17-A release from other cells in response to these T cell specific stimulatory signals seems less likely. Our overall results thereby strongly suggest that neither the leukemia nor the chemotherapy has any selective suppressive effect on T
H17 cells compared with other T cell subsets. Release of IL17-A from other cells than T cells activated in response to these stimulatory signals seems less likely.
Immunosuppressive CD4
+ T
REG cells inhibit effector T cells and NK cells [
36,
37]. Depletion of T
REG cells enhances murine anti-cancer immunity [
38,
39] as well as vaccine-induced anticancer reactivity in humans [
40]. Increased levels of circulating T
REG cells seem to occur in untreated AML [
9,
10], although these previous studies did not analyze FoxP3
+ cells. We observed increased levels of circulating FoxP3
+ T
REG cells in patients with untreated AML. Even though previous studies suggest that at least certain cytotoxic drugs can reduce the levels of circulating T
REG cells [
41], our present results showed that the frequencies of T
REG cells were increased before and following intensive AML chemotherapy. Studies in a second cohort of patients demonstrated that the absolute levels of Treg cells even increased during cytopenia for certain patients. This was true even for patients regenerating after therapy, an observation that is also supported by another study [
9]. The relative levels of CD4
+ CD25
HIGH T
REG cell were also increased in untreated AML and during cytopenia. The variations of various T cell subsets relative to T
REG cells then suggest that the various subsets are differentially affected, and the increased T
REG levels may then be caused either by T
REG proliferation or development from T cell progenitors or other T cell subsets.
We observed a correlation between the levels of TH1 and TC1 cells both for AML patients and healthy controls, and this correlation was maintained during severe therapy-induced cytopenia. In contrast, the correlation between TH1 and TH17 levels was detected only for AML patients with untreated disease and therapy-induced cytopenia. These observations further support the hypothesis that various T cell subsets are differentially affected in AML patients.
The levels of circulating T
H17 cells were significantly higher in males than females for healthy controls, and T
H1 cells were also higher in males for elderly controls. In a recent gene expression study of activated T cells the responsiveness was generally higher in women than in men, but IL17-A was the only effector gene that showed highest expression in men [
42]. Estrogen response elements in the promoter regions of several immune genes could possibly explain the gender differences [
42]. This gender differences reached only borderline statistical significance for patients with untreated AML and could not be detected in our cytopenic patients. Thus, the influence of gender differences on circulating T cell subsets becomes less important after chemotherapy.
Exogenous IL17-A had no or only weak effects on functional AML cell characteristics; even though some differences were statistically significant their biological significance can be questioned. Even though the relative levels of TH17 cells are maintained after chemotherapy, we conclude that TH17 cells may affect AML cells indirectly through their immunoregulatory effects; whereas direct effects of IL17-A on the leukemic cells probably do not have any major impact.
Authors' contributions
The work presented here was carried out in collaboration between all authors. EE and ØB defined the research theme. EE designed methods and experiments, carried out the laboratory experiments, analyzed the data, interpreted the results and wrote the paper. EE, JS and BTG co-designed the phosphoflow experiments, and co-worked on the associated data collection and their interpretation. KL had responsibility of the patients' medical journals, tables, interpretation, and presentation. ØB contributed in all aspects of the study. All authors have contributed to, seen and approved the manuscript.