Altered Development of NKT Cells, γδ T Cells, CD8 T Cells and NK Cells in a PLZF Deficient Patient

In mice, the transcription factor, PLZF, controls the development of effector functions in invariant NKT cells and a subset of NKT cell-like, γδ T cells. Here, we show that in human lymphocytes, in addition to invariant NKT cells, PLZF was also expressed in a large percentage of CD8+ and CD4+ T cells. Furthermore, PLZF was also found to be expressed in all γδ T cells and in all NK cells. Importantly, we show that in a donor lacking functional PLZF, all of these various lymphocyte populations were altered. Therefore, in contrast to mice, PLZF appears to control the development and/or function of a wide variety of human lymphocytes that represent more than 10% of the total PBMCs. Interestingly, the PLZF-expressing CD8+ T cell population was found to be expanded in the peripheral blood of patients with metastatic melanoma but was greatly diminished in patients with autoimmune disease.


Introduction
The promyelocytic leukemia zinc finger (PLZF, ZBTB16) transcriptional regulator is a member of the BTB/POZ-ZF (Broad complex, tramtrack, bric-à-brac or poxvirus and zinc finger-zinc finger; BTB-ZF) family of proteins that have a wide variety of biological activities [1]. Over the last few years, it has become apparent that BTB-ZF proteins are critical regulators of immune system development and function. For example, Bcl6 has been shown to be necessary for both the B cell germinal center reaction [2,3] as well as for the development of follicular helper T cells [4,5]. ThPok has been shown to be necessary and sufficient for CD4 T cell development [6] and the B cell versus T cell commitment step is controlled by LRF (leukemia/lymphoma related factor) [7]. Mazr influences CD8 T cell development, in part by regulating the expression of ThPok [8].
PLZF was first identified due to a t(11;17)(q23;q21) translocation that fused PLZF with RARa(retinoic acid receptor alpha) in some patients with acute promyelocytic leukemia (APL) [9]. The fusion protein that results from this translocation was subsequently shown to be an important determining factor in the development APL and its response to therapy with retinoic acid [10]. The transcription factor has since been implicated in a wide variety of developmental processes, including axial skeletal patterning [11], CNS development [12], male spermatogenesis [13,14], and immune function [15].
Recently it was shown that PLZF is highly expressed in mouse and human invariant natural killer T (iNKT) cells [16,17]. NKT cells represent a subset of CD3 + lymphocytes that serve as a bridge between the innate and adaptive immune systems. Invariant NKT cells represent a subset that expresses a CD1d-restricted Va24Ja18 TCR in humans [18,19]. Both variant and invariant subsets are capable of rapid and copious secretion of a wide variety of Th1 and Th2 cytokines upon initial stimulation, although evidence suggests that the specific roles of each subset may differ [20]. The invariant TCR expressed by iNKT cells recognizes lipid antigens presented in the context of the MHC class I-like molecule CD1d. Invariant NKT cells, which have been extensively studied in mice, are, comparatively, quite rare in humans. For example, in mice 30-40% of T cells in the liver are iNKT cells; in humans only ,1% of hepatic T cells are iNKT cells [21,22,23]. Only 0.008-1.1% of human peripheral blood T cells and only 0.001-0.01% of human thymocytes have been reported to be iNKT cells [24,25,26].
Our lab, and others, have shown that in mice, PLZF controls the development of essentially all of the innate-like features of NKT cells [16,17]. For example, PLZF-deficient NKT cells do not acquire the typical ''activated'' phenotype characterized by high expression of CD44 and CD69 [16,17]. PLZF deficient NKT cells also do not constitutively express granzyme B or the mRNA transcript for IL-4 [16] and fail to acquire the capacity to secrete multiple cytokines upon primary stimulation [16,17]. Furthermore, the frequency of NKT cells is substantially reduced in PLZF-deficient mice and the cells accumulate in the lymph nodes and spleen rather than in the thymus and liver. Overall, the phenotype of PLZF-deficient NKT cells is highly reminiscent of naïve, conventional CD4 T cells [16]. In contrast, ectopic expression of PLZF in conventional T cells results in the acquisition of innate T cell-like characteristics such as an activated phenotype, the rapid secretion of Th1 and Th2 cytokines in response to an initial stimulus and homing to non-lymphoid tissues [17,27,28].
Recent studies have shown that PLZF expression is not strictly limited to invariant NKT cells in mice, but can also be found in a specific subset of cd T cells that express a Vc1.1Vd6.3 TCR [29,30,31]. This subset of ''NKT'' cd T cells functionally resemble invariant NKT cells in that they co-secrete both IFN-c and IL-4 upon primary activation. Importantly, PLZF has been shown to be required for the innate T cell-like characteristics of NKT cd T cells [29]. These studies, together with the findings in NKT cells, highlight an essential and non-redundant role for PLZF in the development of innate T cell effector functions.
In addition to directly controlling the function of the cells it is expressed within, PLZF impacts immune function in trans. Of great interest, studies show that the IL-4 produced by these PLZFexpressing cells profoundly alters the CD8 T cell compartment [32,33]. In mice with an expanded PLZF-expressing T cell compartment, CD8 T cells were found to take on an innate-like phenotype, represented by increased expression of CD44, Eomes and an enhanced capacity to secrete IFN-c. Such mice also harbor increased numbers of germinal center B cells and high serum levels of IgE, in concordance with their heightened Th2 responses [31]. These data show that innate-like T cells, such as NKT cells, have a broad impact on the immune response.
The role of NKT cells in disease is complex and appears to be dependent on both the NKT cell subtype and the microenvironmental context. In mice, NKT cells have been shown to be important in the suppression of solid tumors as a consequence of interactions with dendritic cells and other lymphocytes [34,35,36,37,38,39,40]. In contrast, the immunomodulatory activity of NKT cells can also influence the immune response against autoantigens. For example, autoimmunity in the Type 1 diabetes-susceptible NOD mouse appears to be exacerbated by a deficiency of invariant NKT cells since it is alleviated by NKT activation [41]. In humans with autoimmune diseases, including systemic lupus erythematosus (SLE) [42], diabetes mellitus, and rheumatoid arthritis, decreases in circulating NKTs have, in some studies, been shown to correlate with the frequency of disease [43].
Spurred by recent reports of a broader expression pattern of, and functional relevance for, PLZF in mouse lymphocytes, we examined the full breadth of PLZF expression in human T cells. Furthermore, analysis of the only known person to harbor a biallelic loss of PLZF enabled us, for the first time, to examine the role of this transcription factor in the development of human T cells. We found that in addition to iNKT cells, nearly all cd T cells and natural killer (NK cells) expressed PLZF. Furthermore, more than 5% of CD8 + T cells and ,2% of CD4 + T cells were found to express the transcription factor. Therefore, in total, more than 10% of human PBLs express PLZF. Finally, to study the importance of PLZF for the development of these various cell types, we obtained peripheral blood samples from the only person known to harbor a biallelic loss of functional PLZF [44]. Overall, our data suggest that differences in PLZF expression represent a significant divergence between the mouse and human immune system.

Blood Draw
Blood draw protocols and ethics were reviewed and approved by the Internal Review Board's (IRB) at Memorial Sloan-Kettering Cancer Center (New York, NY), the Hospital for Special Surgery (New York, NY), the Winthrop-University Hospital (Mineola, NY) and Universitaetsklinikum Essen (Essen, Germany). Written informed consent forms were obtained from all donors.

Cytokine Expression
T cells were plated in X-vivo media with PMA (100 ng/ml) and ionomycin (500 ng/ml) or beads coated with anti-CD3/CD28 (Miltenyi Biotec). Brefeldin A was added for the last five hours of a 6 hour culture. Cells were stained for surface markers, fixed, made permeable and stained with anti-IFNc AlexaFluor 488 (BD). NK cells were activated with beads coated with anti-NKp46 and anti-CD2 (Miltenyi Biotec). Cytokines were analyzed with cytokine bead arrays (BD).

Patient Samples
Patients with an established diagnosis of systemic lupus erythematosus (SLE, n = 6), primary Sjögren's syndrome (n = 3), or metastatic melanoma (n = 10) and healthy controls (n = 8) were consented in accordance with local institutional review board guidelines. Lymphocytes were isolated from patients with stage III-IV melanoma by leukapheresis. Lymphocytes from fresh peripheral blood for other samples were isolated by Ficoll extraction.

Real Time PCR
Sorted cells were suspended in 1 ml Trizol and frozen at -80 degrees. RNA was isolated using Qiagen's RNeasy mini kit. Following reverse transcription, the abundance of PLZF cDNA was quantified by real-time PCR using the MasterMix (2X) HotStart-IT Syber kit (USB). Expression levels were normalized against the housekeeping gene GAPDH, according to the formula: relative PLZF induction = 2 CT(GAPDH) -CT(PLZF) .

Statistical Methods
Statistical analysis was done either with a two-tailed, unpaired Student's t-test or the Mann-Whitney U-test, which is a more robust for small sample sizes.

PLZF expression in human T cells
PLZF is expressed both in mouse and human invariant NKT cells [16]. Recent studies in mice, however, show that PLZF expression is not limited to this cell type [29,30,31]. Therefore, we more carefully examined the expression of PLZF in PBLs from healthy volunteers by FACS. Similar to mice, PLZF was not detected in ''conventional'' T cells (Fig. 1A). The cd TCR and iNKT cell-negative, CD3 + T-cell compartment can be demarcated by the expression of CD161 and CD56. These cells are found at fairly high frequencies, with CD56 + T cells on average at 1.7% (+/-2.5%) and CD161 + T cells often found at more than 10% of the total T cells (Fig. 1B). CD56 + T cells, which are often are considered to be non-invariant NKT cells, did not express PLZF (Fig. 1C). Similarly, CD161 lo cells (white box, Fig. 1B) also did not express PLZF. Of particular interest, however, was a distinct population of T cells (3.4% +/22.5%) marked by high levels of CD161 ( Fig. 1B) that clearly expressed PLZF (Fig. 1C).
The majority of the PLZF-expressing CD161 hi T cells were found to be CD8 + (Fig. 1D). Furthermore, the CD161 hi CD8 + T cells (Fig. 1E), expressed higher levels of PLZF as compared to the CD161 hi CD4 + T cells (Fig. 1F). PLZF expression was lower than in iNKT cells (Fig. 1F). PLZF expression levels were consistent among cells from eight healthy donors (Fig. 1G). Expression levels were also confirmed by rtPCR of sorted cell populations (Fig. 1H). The CD161 hi CD8 + population represented, on average, ,6% of the total CD8 + T cells (Figs. 1B, 1E). PLZF is not expressed in any identified CD8 + T cell population in mice [16]and unpublished data). Therefore, PLZF expression defines an abundant T cell subgroup in humans that does not exist in mouse peripheral blood.
To determine if the CD161 hi subset arose from the CD161 2 population as a result of activation, CD8 + and CD4 + T cells that were CD161 2 (Fig. 1I, black lines) or CD161 hi (Fig. 1I, red lines) were sorted and cultured with beads coated with antibodies against CD3 and CD28. After three days, low levels of CD161 were induced on the CD161 2 T cells (Fig. 1J, black lines). CD161 hi CD8 + T cells were found to retain high levels of CD161 expression post-activation (Fig. 1J, red lines). The CD161 hi CD4 + T cells, however, downregulated CD161 post-activation (Fig. 1J, red lines). Therefore, the CD161hi phenotype is specific to the PLZFexpressing CD8 + T cells and expression of high levels of CD161 cannot be induced by activation. Furthermore, PLZF was not induced in non-PLZF expressing T cells by activation (data not shown).
The frequency of PLZF-expressing CD161 hi CD8 + T cells is altered in patients with autoimmunity or advanced melanoma To begin to study the role of PLZF-expressing CD161 hi CD8 + T cells in the human immune response, we obtained peripheral blood samples from patients with systemic autoimmune diseases or malignancy. Patient samples were collected based only on availability; no selection criteria were utilized and treatment status was unknown. Monocyte-depleted PBMCs were stained with antibodies against CD3, CD4, CD8, CD161, CD56 and cd TCR, then analyzed by flow cytometry. Two representative samples acquired from patients diagnosed with systemic lupus erythematosus (SLE) is shown in Fig. 2A. CD161 hi T cells appeared to be reduced, but not to a level that reached statistical significance, when the SLE samples were compared to healthy donors ( Fig. 2A). The apparent reduced frequency of CD56 + T cells also did not reach statistical significance (data not shown). There was, however, a clear reduction in the percentage of CD8 + cells within the CD161 hi population in patients with SLE (Fig. 2B). Similarly, the frequency of CD161 hi cells within the CD3 + CD8 + T cell population was significantly reduced in the SLE samples, as compared to healthy controls (Fig. 2C). Again, the apparent reduction of CD3 + CD8 + CD56 + was not statistically significant (data not shown). A similar trend was seen in a limited survey of patients with primary Sjogren's syndrome (data not shown). Finally, the residual CD161 hi CD8 + T cells in the SLE patients expressed little or no PLZF (Fig. 2D).
Next, we evaluated the frequency of CD161 + T-cells in blood samples from ten patients with advanced melanoma. In contrast to our findings in SLE and Sjogren's patients, we observed a substantial increase in the frequency of CD161 hi T cells in the melanoma cohort as compared to healthy controls (Fig. 3A). Different than the patients with autoimmunity, there was no change in the frequency of the CD8 to CD4 ratio within the CD161 hi T cells (Fig. 3B). Among the CD8 + T cells, however, there was a substantial increase in the percentage of CD161 hi T cells (Fig. 3C). Also different than the SLE samples, the CD161 hi CD8 + T cells in the patient samples were found to express PLZF at levels similar to healthy donors (Fig. 3D). These data, along with published studies [45], show that the PLZFexpressing PBL T cell population changes in patients with disease.

T cell subset deficiencies as a consequence of a biallelic loss of functional PLZF
Recently, the first and, to our knowledge, only person with a biallelic loss of functional PLZF was identified [44]. The maternal allele of PLZF was found to have a missense mutation that resulted in a methionine to valine substitution within a zinc finger domain. This mutation is paired with the paternal chromosome harboring a de novo deletion of ,8 Mbp, resulting in the loss of the gene encoding PLZF. Peripheral blood from this person was obtained and analyzed by FACS. In the absence of PLZF, the total percentage of CD3 + T cells among PBLs and the percentage of CD4 + and CD8 + T cells were similar to healthy controls (Fig. 4A). The frequency of the CD56 + and the CD161 hi T cell populations, however, were reduced as compared to the mean frequency found in healthy controls (Fig. 4B,C) and the few CD56 + and CD161 hi cells found did not express PLZF (Fig. 4D). The few CD161 hi cells observed in this donor were strongly skewed towards CD4 rather than CD8 expression, which contrasted with the pattern observed in healthy controls (Fig. 4E,F). Additionally, there was a decrease in frequency of CD161 hi DN T cells, as compared to controls (Fig. 4E,F).

Human iNKT cells require PLZF for normal development
In mice, iNKT cells develop in the absence of PLZF, but they do not acquire innate T cell characteristics [16,17]. For example, in mice, PLZF-deficient iNKT cells accumulate in the lymph nodes rather than in the liver, do not express high levels of CD44 and CD69 and, also, fail to make cytokines upon primary activation. PLZF-deficient mouse iNKT cells are highly skewed towards CD4 and express low levels of NK1.1 [16]. iNKT cells are also present in the peripheral blood of PLZF-deficient mice, but at a reduced frequency as compared to wild type mice (data not shown). The frequency of human iNKT cells in the PLZF-deficient donor was only slightly reduced as compared to the mean frequency found in healthy controls (Fig. 5A). However, the phenotype of the PLZFdeficient iNKT cells clearly resembled that of iNKT cells from PLZF-deficient mice. In particular, PLZF-deficient human iNKT cells were strongly skewed towards expressing CD4 at the expense of the DN population (Fig. 5B). Furthermore, the PLZF deficient iNKT cells were largely CD161 (NK1.1) low (Fig. 5C). These data suggest that, similar to the findings in mice, PLZF impacts the development of iNKT cells in humans.

Human cd T cells express PLZF
The reduction in the frequency of CD161 hi DN T cells in the absence of PLZF (Fig. 4E,F) prompted us to more carefully examine this population of T cells. A large percentage of the CD161 hi DN T cells proved to be cd T cells (data not shown). We, and others, recently reported the existence of a minor subset of NKT cell-like cd T cells in mice that express high levels of PLZF [30,31]. Unlike the subset-restricted pattern of PLZF expression in mouse cd T cells, all human peripheral blood cd T cells were found to express PLZF (Fig. 6A). The expression of PLZF was confirmed by comparing cd T cells from healthy donors to cd T cells from the PLZF-deficient donor (Fig. 6A). cd T cells were found to develop in the absence of PLZF and were present at a frequency similar to that published by others (,4%) [46] and to a larger panel of samples we analyzed (Fig. 6B and data not  shown). The data suggests that the frequency of CD161 hi cd T cells was reduced in the absence of PLZF, while the frequency of CD8 + cd T cells was increased (Fig. 6B), however, cd T cell phenotypes were found to be highly variable among healthy donors.

Human NK cells express PLZF
Less than 5% of mouse NK cells constitutively express PLZF [16] and PLZF is not induced in mouse NK cells by activation (data not shown). Furthermore, mouse NK cells develop and appear to function normally in PLZF-deficient mice ( [16] and data not shown). Therefore it was very surprising to find that nearly all  human NK cells expressed the transcription factor. The expression of PLZF was made clear by comparing samples from healthy donors to the PLZF-deficient donor (Fig. 7A) PLZF-deficient NK cells were found to express lower levels of CD56, CD16 and CD161 (Fig. 7B). To assess the functional relevance of PLZF in human NK cells, NK cells from healthy controls and the PLZF-deficient donor were activated with beads coated with antibodies against CD335 (NKp46) and CD2. Supernatants were collected after three days for cytokine analysis. There was a clear increase in the amount of secreted IFN-c, TNF and Granzyme B in the PLZF-deficient sample as compared to the control samples (Fig. 7C). GM-CSF was also found to be increased, but to a lesser degree. Therefore, PLZF expression appears to subdue cytokine secretion by human NK cells.

Discussion
We show that in people, CD8 + and CD4 + CD161 hi T cells, all cd T cells, NKT cells and all NK cells express PLZF. Collectively, these populations represent more than 10% of the total human peripheral blood lymphocytes. In contrast, PLZF expression in mice is limited to NKT cells and a minor subset of cd T cells, which combined, account for ,1% of the total lymphocytes. Furthermore, few, if any, mouse NK cells express PLZF. Therefore, PLZF arguably plays a far more significant role in the human immune system than in that of the mouse.  Although the frequency of cd T cell in PLZF deficient sample appears higher than the controls, the percentage is actually consistent with a broader analysis (data not shown) and with published reports. The bottom panels show CD161 and CD8 expression on cd T cells from the healthy control and the PLZF deficient donor. cd T cells were identified as shown in the top panels. Data from PLZF deficient donor were consistent from three independent blood samples. Numbers indicate the frequency of cells within each panel. doi:10.1371/journal.pone.0024441.g006 The frequency, phenotype and/or function of all of the PLZFexpressing cell types were altered in some way in samples obtained from a person with a biallelic loss of functional PLZF. For example, the phenotype of CD1d-restricted NKT cells was altered and the frequency of CD161 hi CD8 + T cells was reduced, suggesting that these T cells are dependent upon PLZF for development. cd T cells were present in the PLZF-deficient donor, but appeared to have altered CD161 and CD8 expression. In the absence of PLZF, NK cells expressed lower levels of CD56 and the activating receptors CD16 and CD161. Surprisingly, we found that PLZF-deficient NK cells produced substantially greater amounts of several different cytokines. No obvious alterations in the PLZF-expressing cell types were detected in a limited analysis of the donor's mother, who carries a point mutation in one copy ZBTB16 (data not shown). As mentioned above, the donor's father carries two non-mutated copies of the PLZF gene and, therefore, blood samples were not obtained. Finally, the donor has not shown overt signs of autoimmunity nor increased susceptibility to infectious disease. Indeed, although the donor is EBV positive, he does not display signs of active disease.
Rather surprisingly, we found that PLZF was not expressed in CD3 + CD56 + T cells. Nevertheless, the frequency of CD3 + CD56 + T cells was reduced in the PLZF-deficient donor, implicating PLZF in the development of these cells. It is possible that expression of PLZF in CD3 + CD56 + T cells may occur during T cell development, when it directs the acquisition of an innate Tcell phenotype in this population, but is then downregulated prior to the migration of these cells into the periphery, similar to what was found for embryonic CD4 T cells [47]. Alternatively, the reduced frequency of CD56 + T cells may be due to cell extrinsic factors, such as altered cytokine expression by other PLZFexpressing cells.
Recently published data suggests that a large percentage of CD161 hi CD8 + and CD161 hi double negative T cells express an invariant Va7.2 TCRa chain [48]. T cells expressing this invariant TCR belong to a subset known as mucosal-associated invariant T cells (MAIT cells). The finding that human, but not mouse, MAIT cells express PLZF suggests a functional divergence in this T cell subset.
To begin to understand the role of PLZF-expressing CD161 hi CD8 + T cell lineage in the human immune system, we analyzed the frequency of this population in blood obtained from patients with either autoimmunity or cancer. Patients with SLE were found to have a highly significant decrease in the percentage of CD8 + T cells within the CD161 hi subpopulation. Similar results were obtained from a small cohort of patients with primary Sjögren's syndrome. Interestingly, patients with advanced melanoma were found to have a significant increase in the frequency of CD161 hi T cells. In one extreme example, ,16% of the total T cells were CD161 hi and, in particular, nearly 40% of the patient's CD8 T cells were CD161 hi . While these data have not been correlated with clinical outcomes, it is clear that the PLZFexpressing population is dynamic and that it changes with different types of immunological stress.
The patient data are consistent with a previous report showing that CD161 hi CD8 + T cells are reduced in patients with a variety of rheumatic diseases including SLE, MCTD (mixed connectivetissue disease), SSc (systemic sclerosis) and PM/DM (polymyositis/ dermatomyositis) [45]. Our data are also consistent with a report showing that CD161 hi CD8 + T cells do not produce IFN-c, TNF-a or IL-2 following TCR activation [49]. A more recent paper suggests that CD161 hi CD8 + T cells contain an IL-17 producing subset as well as IL-22 producing cells [50]. Interestingly, it was found that the IL-17 producing cells were enriched in the livers of patients during chronic hepatitis. The PLZF-expressing CD161 hi CD8 + T cells also express high levels of the IL-18 receptor a chain (data not shown). Therefore it is possible that the cells we study here are the same or similar to those described as having the capacity to efflux and, therefore, survive treatments involving chemotherapy drugs [51]. It was suggested that the Supernatants were collected after 72 hours and analyzed by cytokine bead array. Control-1 was obtained from a healthy donor the day of the sort; control-2 was obtained from a healthy donor and the sample was shipped with the PLZF-deficient sample. FACS analysis was consistent for eight healthy controls. Samples from the PLZF-deficient donor were collected three times and were analyzed in three independent FACS experiments. Cells for the cytokine release experiment were collected from one sample. Sorted cells were split into multiple wells (between 3 and 8 wells) and treated as independent samples for statistical purposes. (** P value = 0.0080; *** P value ,0.0001). doi:10.1371/journal.pone.0024441.g007 CD161 hi T cells are actually precursors of CD161 lo cells and, thus, might represent self-renewing stem cell-like memory CD8 T cells [51]. It will be interesting to determine if these proposed functions are dependent on PLZF.
Overall, we have found that PLZF is expressed in a greater variety of lymphocytes and in a higher percentage of lymphocytes in humans as compared to mice. The function of PLZF-expressing cells appears to be diverse. For example, CD161 hi CD8 + T cells clearly do not make abundant levels of cytokines following activation. Furthermore, PLZF expression in human NK cells appears to limit cytokine production. Therefore, in sharp contrast to other PLZF-expressing lymphocytes, PLZF expression in human CD8 + T cells and NK cells appears to inhibit cytokine expression.