SEARCH

SEARCH BY CITATION

Keywords:

  • Cowden disease;
  • flow cytometry;
  • interferon-gamma

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References

Identification of immune modifiers of inherited cancer syndromes may provide a rationale for preventive therapy. Cowden disease (CD) is a genetically heterogeneous inherited cancer syndrome that arises predominantly from germline phosphatase and tensin homologue deleted on chromosome 10 (PTEN) mutation and increased phosphoinositide 3-kinase/mammalian target of rapamycin (PI3K/mTOR) signalling. However, many patients with classic CD diagnostic features are mutation-negative for PTEN (PTEN M-Neg). Interferon (IFN)-γ can modulate the PI3K/mTOR pathway, but its association with PTEN M-Neg CD remains unclear. This study assessed IFN-γ secretion by multi-colour flow cytometry in a CD kindred that was mutation-negative for PTEN and other known susceptibility genes. Because IFN-γ responses may be regulated by killer cell immunoglobulin-like receptors (KIR) and respective human leucocyte antigen (HLA) ligands, KIR/HLA genotypes were also assessed. Activating treatments induced greater IFN-γ secretion in PTEN M-Neg CD peripheral blood lymphocytes versus healthy controls. Increased frequency of activating KIR genes, potentially activating KIR/HLA compound genotypes and reduced frequency of inhibitory genotypes, were found in the PTEN M-Neg CD kindred. Differences of IFN-γ secretion were observed among PTEN M-Neg CD patients with distinct KIR/HLA compound genotypes. Taken together, these findings show enhanced lymphocyte secretion of IFN-γ that may influence the PI3K/mTOR CD causal molecular pathway in a PTEN mutation-negative CD kindred.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References

Germline phosphatase and tensin homologue deleted on chromosome 10 (PTEN) mutation is causally implicated in 80% Cowden's disease (CD), an autosomal dominant Hamartoma syndrome characterized by macrocephaly, benign thyroid or breast disease and increased cancer risk [1]. Loss of PTEN tumour suppressor activity up-regulates PI3K/mTOR signalling implicated in the pathobiology of Hamartoma syndromes [2,3]. Interestingly, PI3K/mTOR also transduces mitogenic signals in response to cytokines [4]. Many CD patients are mutation-negative for PTEN (PTEN M-Neg) [1], although a small subset may be associated with germline succinate dehydrogenase (SDHB, SDHD) gene mutations that also deregulate PI3K/mTOR signalling [5].

Clues to the identity of unknown immune or genetic modifiers of inherited cancer syndromes may be provided by murine model studies. The onset, disease spectrum and progression of CD arising from a specific exon 5 PTEN mutation are profoundly affected by immunogenetic murine strain differences [6] that involve interferon (IFN)-γ responses [7] and natural killer (NK) cell receptor expression [8]. IFN-γ produced by activated NK or lymphoid cells may deregulate the PI3K/mTOR pathway by activation of protein kinase B (AKT) and extracellular-regulated kinase (ERK) [9,10]. These signals promote functional dissociation of the TSC1/2 complex and mTOR hyperactivity [11,12], implicated in the development of CD hamartomas and cancer [2,3].

Secretion of IFN-γ can be modulated by killer cell immunoglobulin-like receptors (KIR) [13], a family of 16 genes and/or pseudogenes that interact with cognate human leucocyte antigen (HLA) ligands [14]. Long-tailed KIR genes (KIRL) typically have an inhibitory function, while short-tailed KIR (KIRS) are activating receptors [14], although KIR2DL4 may be either activating or inhibitory [15]. The KIR locus is strictly co-ordinated, and haplotypes segregate into two groups. KIR group A haplotypes include genes encoding inhibitory receptors for each of the four KIR ligands as well as KIR2DL4, KIR3DL3 and a single activating gene, KIR2DS4. Conversely, group B haplotypes carry additional activating KIRs [16]. Individuals can be categorized as having A/A or B/x genotypes, which are homozygous for group A haplotypes or heterozygous or homozygous for group B haplotypes, respectively [16]. Activity of individual receptors may be influenced by KIR/HLA compound genotypes. HLA-A is a weak ligand for KIR3DL2 [17], while associations of HLA-B-KIR3DL1 [18] and HLA-C-KIR2DL1 and HLA-C-KIR2DL2/3 [19] are well recognized. KIR2DL2, KIR2DL3 and activating KIR2DS2 each have two immunoglobulin (Ig)-like domains. The inhibitory alleles bind HLA-C group 1 (HLA-C1) alleles, encoding ser77/asp80 of HLA-Cw a1 while the activating KIR2DS2 could also potentially bind these alleles. In contrast, the HLA-C group 2 (HLA-C2) alleles, encoding asp77/lys80, bind to KIR2DL1 and potentially to KIR2DS1 [20]. Among the KIRs with three Ig-like domains, KIR3DL1 recognizes HLA-B molecules sharing the Bw4 motif [21]. KIR/HLA genetics regulate lymphocyte activation thresholds and cytokine repertoires [21]. Associations between KIR/HLA immunogenetics, IFN-γ responses and PTEN M-Neg CD, however, remain unclear.

This study has assessed IFN-γ responses and KIR/HLA compound genotypes in a CD kindred that met International Cowden's Consortium (ICC) criteria [1] but was mutation-negative for PTEN. Further screening was also negative for succinate hydrogenase SDHB, SDHD CD susceptibility genes [5]. Screening for LKB1, which may interact with PTEN [22] and is causally implicated in Peutz–Jeghers disease, a Hamartoma-syndrome with some CD-like features [23], was also negative. Here we report greater secretion of IFN-γ in PTEN M-Neg CD peripheral blood lymphocytes (PBLs) in response to activating treatment versus healthy control PBLs. Increased frequency of activating KIR genes and haplotypes and increased potentially activating KIR/HLA compound genotypes, together with reduced frequency of inhibitory KIRs, were also found in the PTEN M-Neg CD kindred.

Patients and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References

Study populations

The study kindred of CD patients manifests an autosomal-dominant syndrome with macrocephaly, as well as benign and malignant breast, thyroid and other neoplasms. The proband was diagnosed with macrocephaly, breast and thyroid cancer before the age of 40 years and had a strong family history of macrocephaly, breast and other cancers (Fig. 1). Because the index case (Fig. 1; arrowed) had three major criteria, including macrocephaly and other family members, manifesting at least one single major criterion with or without minor criteria, the kindred was assessed externally and was considered to have PTEN mutation-negative CD [Dr Charis Eng, Dana Human Cancer Genetics Unit, Dana-Farber Cancer Institute, Boston, MA (now based at the Cleveland Clinic Main Campus, Cleveland, OH) personal communication to P. J. M.]. Written consent for the study was obtained from all the patients and the study was granted ethical approval by ORECNI (the Office of Research Ethics Committees, Northern Ireland). Detailed mutation screening of the entire PTEN gene and promoter region was carried out by Dr Eng in 11 surviving members of the pedigree. Comprehensive mutation analysis of all nine coding exons, flanking intronic regions and minimal promoter regions of PTEN and examination for large deletions and rearrangements were conducted as described previously [24]. No mutations were found. Microsatellite linkage analysis to chromosome 10q was performed, but linkage was excluded. Testing for other genes included sequence and multiplex ligation-dependent probe amplification (MLPA) analysis of SDHB, SDHD and LKB1 genes. All these genes were normal. All patients and most healthy controls for KIR/HLA genotyping (n = 200) and PBL/NK cell functional studies are Northern Irish Caucasians [25].

image

Figure 1. Pedigree of phosphatase and tensin homologue deleted on chromosome 10 (PTEN) mutation-negative CD that fits consensus diagnostic criteria for Cowden's disease. Members with cancer are shown in full shading, those with benign tumours with a central dot. Ages at which cancer appeared are given (see Results for details). The study proband is indicated by the arrow.

Download figure to PowerPoint

Lymphocyte and NK cell functional studies

These studies were conducted in five PTEN M-Neg CD patients and five healthy age- and sex-matched controls. Peripheral venous blood (20 ml) was collected in heparinized tubes from CD patients and controls, transferred to cell culture flasks and diluted with one volume of Iscove's modified Eagle's medium (IMDM) medium (Invitrogen, Paisley, PA49RF, UK) containing 10% human serum (Promocell, Heidelberg, Germany), 1% non-essential amino-acids (Invitrogen) and 1% sodium pyruvate. Samples were incubated overnight at 36°C in a humidified 5% CO2 incubator. All the flow cytometry analysis was performed using a Becton Dickinson (San Jose, CA, USA) LSR2 flow cytometer and cell sorting was performed using a Becton Dickinson fluorescence activated cell sorter (FACSAria II), prior to flow cytometry analysis.

NK cell assays

In working with NK cells, we used the established definition of CD56e+ and/or CD16e+ and CD3- lymphocytic cells within a light-scatter gate for lymphocytes [26]. Phenotypic analysis, sorting and isolation of NK cells were carried out using allophycocyanin (APC)-conjugated anti-CD56, APC-conjugated anti-CD16 and peridinin chlorophyll cyanin (PerCP Cy) 5·5-conjugated anti-CD3 monoclonal antibodies (mAbs). To determine NK cell numbers, 100 µl diluted blood for each sample were incubated with labelled mAbs within a Trucount tube for 20 min at room temperature, then lysed with NH4Cl. The Trucount tubes contained a calibrated number of dye-labelled beads. NK cell numbers were calculated as {NK cells/µl sample = [(CD56+/CD3 cell count)/(bead count)] × bead concentration/blood volume (µl) × blood dilution factor}. To isolate the NK cell population, we initially sorted on the FACSAria II by the NK cell gating strategy (CD56+, Cd16+ and CD3), as outlined above, then sorted directly into culture plates. Cells were sorted into ionomycin (1 µg/ml) or Cytostim (20 µl/ml) versus control CellGro (CellGenix, Freiburg, Germany) Dulbecco's modified Eagle's serum-free medium (DMEM) for stimulation studies. Cytostim is a proprietary antibody-based reagent whose composition is protected commercially, and molecular weights of active constituents are unavailable.

Assay of IFN-γ secretion

IFN-γ secretion was assayed in PBLs treated by ionomycin (1 µg/ml) or Cytostim (20 µl/ml cell suspension) (Miltenyi Biotec, Bergisch Gladbach, Germany) on separate occasions using magnetic affinity cell sorting (MACS) IFN-γ secretion assay detection kits (Miltenyi Biotec), according to the manufacturer's instructions. Briefly, diluted blood samples were incubated in erythrocyte lysing solution for 10 min, washed in cold buffer, centrifuged and the resulting pellet resuspended in medium. Cells were plated at 5 × 106 cells/well in multi-well plates and stimulated by ionomycin (1 µg/ml) or Cytostim (20 µl/ml). T cell receptor (TCR)-dependent and -independent pathways may be involved in T cell activation and production of IFN-γ[27]. We used appropriate stimulants to probe these pathways in the present study. Cytostim acts similarly to a superantigen and activates T cells by binding to the TCR [28]. Conversely, ionomycin is a calcium ionophore and activator of protein kinase C that has been used widely to model or bypass early receptor-mediated signalling [29]. We used Cytostim concentrations according to the manufacturer's instructions [28] and ionomycin exposure that has been shown previously to induce T cell activation or IFN-γ secretion [30]. Cytokine catch reagent (Miltenyi Biotec) was added (20 µl to a final volume of 100 µl), samples were incubated on ice, diluted in 5 ml warm medium and placed in a rotating incubator for 45 min at 37°C. Samples were then washed in buffer, centrifuged and pellets resuspended in cold buffer containing 20 µl FITC-labelled anti-IFN-γ detection antibody (Miltenyi Biotec). Prior to flow cytometry analysis, 7-amino-actinomycin (7-AAD) was added to exclude dead cells from the analysis. 7-AAD intercalates into double-stranded nucleic acids. It is excluded by viable cells but can penetrate cell membranes of dying or dead cells [31]. The percentage of IFN-γ-expressing cells was assayed using a Becton Dickinson LSR2 flow cytometer.

Assay of NK cell IFN-γ secretion

Total NK cells were gated using labelled mAbs. Cells were suspended in CellGro medium at 50 000 cells per well, then placed in multi-well plates containing control medium only or ionomycin (1 µg/ml). After 2 h, samples were then incubated with APC-labelled anti-IFN-γ, as outlined above, for a further 2 h at 37°C. IFN-γ secretion was quantified as outlined above.

KIR and HLA genotyping

The presence of each KIR gene was initially evaluated by polymerase chain reaction (PCR)-sequence-specific oligonucleotide probe (SSOP) at the Regional Histocompatibility and Immunogenetics Laboratory, City Hospital, Belfast, as described previously [32]. Coding sequences for alleles were taken from the KIR Immuno Polymorphism Database IPD, release 1·4.0, June 2007 [33,34], and additional intronic sequences from European Molecular Biology Unit–European Bioinformatics Institute (EMBL-EBI) (http://www.ebi.ac.uk/ebisearch/) (e.g. Accession numbers AF217486 and AF217486 for KIR2DL2 and EF102434 for KIR2DL5A). Primer sequences were assessed on BLASTN (http://ncbi.nlm.nih.gov/blast) and a KIR allele-specific sequence search engine supplied by IPD (http://www.ebi.ac.uk/ipd/KIR/.html) to ensure their homology with the gene being identified. Luminex-based typing was used to characterize HLA-B and -C alleles [33,34]. KIR haplotypes were compiled from KIR gene-specific and allele-specific typing and assigned as AA, AB or BB, according to those defined previously from haplotype segregation in families [35]. Combinations of KIR and HLA alleles were defined according to known ligand-receptor interactions that influence KIR activity [26,36]

Data analysis

Phenotype frequencies were determined by direct counting of individuals who were positive for a particular KIR specificity. Gene frequencies were calculated using the equation P = 1-√1–f (P = gene frequency and f = phenotype frequency). Effects of disease status or treatment on lymphocyte and NK cell numbers and IFN-γ secretion were assessed by one- or two-way analysis of variance (anova) with the Tukey post-hoc test. Descriptive statistics include the mean and standard error (mean ± s.e.) (PASW statistics, release 18 2009).

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References

Patients

The four generation family pedigree is shown in Fig. 1. The proband (III.5; arrowed) had breast and thyroid cancer and macrocephaly. Members I.2 had breast cancer, I.4 had leukaemia, II.1 and II.3 had colon cancer while II.4 has colonic polyps. II.2 had cancer of the eye and died aged 7 years. II.7 had oesophageal cancer, although was a smoker; III.7 and IV.2 had melanomas and III.2 had cutaneous common moles, III.3 has macrocephaly, III.4 had a spinal ependymoma and macrocephaly, III.6 has a parathyroid adenoma, III.7 had benign thyroid disease and malignant melanoma, III.8 has benign breast lumps, III.13 had benign thyroid disease, III.14 has Hodgkin's disease and IV.1 has macrocephaly with ventricular enlargement.

Lymphocyte/NK cell function studies

These assays were carried out in five PTEN M-Neg CD patients (members III.8, III.5, III.7, III.4 and III.13 representing CD patients 1–5, respectively) whose specific KIR and HLA gene expression and compound KIR/HLA genotypes are shown (Tables 1–3) and five healthy controls. NK cell numbers were lower in CD patients versus healthy controls (PTEN M-Neg CD = 1531 ± 339·2 versus 2022 ± 339·2 control; P = 0·046; anova), although no differences of total unfractionated peripheral blood lymphocytes (PBLs) were observed. NK and lymphocyte numbers were not affected significantly by ionomycin treatment in CD patients or controls. The percentage of PTEN M-Neg CD PBLs that secreted IFN-γ increased from 3·38 ± 1·1% to 5·48 ± 1·1% versus 2·38 ± 1·1% to 2·86 ± 1·1% in controls, after ionomycin treatment. PTEN M-Neg CD PBLs also showed greater IFN-γ fold induction (CD = 3·0 ± 0·41 versus 1·5 ± 0·4 controls; P = 0·042; anova) and greater post-treatment IFN-γ fluorescence intensity than healthy controls [7148 ± 587 (CD) versus 4339 ± 587 (control); P < 0·001; anova; Fig. 2a and b]. Interactive effects of disease category (CD versus control) and ionomycin treatment on IFN-γ fluorescence intensity were significant; P = 0·006; two-way anova). IFN-γ fluorescence intensity increased in PTEN M-Neg CD versus control NK cells from 3270 ± 782 to 4500 ± 782 (CD) versus 3187 ± 782 to 3902 ± 782 (control), although differences were not significant.

Table 1.  Killer cell immunoglobulin-like receptor (KIR) gene profiles in the Cowden disease (CD) pedigree.
AllelesInhibitory KIRPseudo GActivating KIR
2DL12DL22DL32DL42DL53DL13DL23DL32DP13DP12DS12DS22DS32DS42DS53DS1NDDEL
% Cont(96)(51)(93)(100)(44)(95)(100)(100)(100)(100)(34)(52)(28)(95)(26)(35)
Patient
  1. Numbers in parentheses represent the frequencies for individual KIR genes in the control population.

III.2111110101111001100
III.4 *CD4111111111111011101
III.5 *CD2111111111111111101
III.6111110111111001100
III.7 *CD3111110111111001100
III.8 *CD1111110111111001100
III.9111110111111001100
III.10111110111111001100
III.11111110111111101100
III.13 *CD5111111111111111101
III.14111110111111101100
Table 2.  Human leucocyte antigen (HLA) genotypes in the phosphatase and tensin homologue deleted on chromosome 10 (PTEN) mutation-negative (M-Neg) Cowden disease (CD) kindred.
PtHLA genotypes
HLA-BHLA-C
AllelesEpitopesAllelesEpitopes
  1. HLA-B and -C allele frequencies are shown. Note that B27 and B44 have the Bw4 epitope while B07 and B40 have the Bw6 epitope. C07 has the C1 epitope while C02, C05 and C15 have the C2 epitope.

III.2B*2705B*4402 Bw4C*0202C*0501 C2
III.4 *CD4B*4402B*4002Bw6Bw4C*0501C*1502 C2
III.5 *CD2B*4402B*4002Bw6Bw4C*0501C*1502 C2
III.6B*2705B*4402 Bw4C*0202C*0501 C2
III.7 *CD3B*0702B*4002Bw6 C*0702C*1502C1C2
III.8 *CD1B*0702B*2705Bw6Bw4C*0702C*0202C1C2
III.9B*4402B*4002Bw6Bw4C*0501C*1512 C2
III.10B*2705B*4402 Bw4C*0202C*0501 C2
III.11B*4402B*4002Bw6Bw4C*0501C*1502 C2
III.13 *CD5B*0702B*4002Bw6 C*0501C*1502C1C2
III.14B*0702B*4002Bw6 C*0702C*1502C1C2
Table 3.  Killer cell immunoglobulin-like receptors/human leucocyte antigen (KIR/HLA) compound genotypes in Cowden disease (CD) patients versus healthy controls.
KIR/HLA genotypeCD%Control%
2DL1-C210051
2DL2-C13743
2DL3-C13781
3DL1-Bw4054
2DS1-C210026
2DS2-C13737
3DS1-Bw47327
image

Figure 2. (a) Phosphatase and tensin homologue deleted on chromosome 10 (PTEN) mutation-negative (M-Neg) Cowden disease (CD) versus control peripheral blood lymphocyte (PBL) interferon (IFN)-γ secretion. PTEN M-Neg CD PBLs showed greater IFN-γ fluorescence intensity after ionomycin treatment than control (P < 0·001; analysis of variance). (b) Illustrative example of flow cytometry assay of PBL IFN-γ secretion in unstimulated or (ionomycin) stimulated conditions from PTEN M-Neg CD patient 4 (specimen 001–9) and a representative healthy control (specimen 001–5) (left and right upper and lower panels, respectively). Dead lymphocytic cells were excluded by use of 7-amino-actinomycin (7-AAD) and lymphocyte gating. IFN-γ secretion is assayed using fluorescence activated cell sorter-labelled anti-IFN-γ detection antibody.

Download figure to PowerPoint

KIR allele frequencies

KIR genotypes were determined in CD patients and compared against a control group (n = 200) from the same population (available at http://www.allelefrequencies.net). All members of the PTEN M-Neg CD kindred showed high frequency for most of the activating KIR genes (KIR2DS1, KIR2DS2, KIR2DS5 and KIR3DS1) but lower frequencies of KIR2DS4 (Table 1). However, the three patients with KIR2DS4 were found to carry only deleted alleles of KIR2DS4, which are non-expressed forms. Hence, all PTEN M-Neg CD patients lacked this gene product expressed at cell surfaces. KIR3DL1 was detected in three of 11 CD patients (Table 1). Allelic subtyping of KIR3DL1 in these three patients was found to be homozygous for the non-expressed allele KIR3DL1*004. Hence, all PTEN M-Neg CD patients lacked functional KIR3DL1 inhibitory receptors at cell surfaces. In contrast, functional KIR3DL1 genes are expressed by 95% of the control population (http://www.allelefrequencies.net).

KIR haplotype profiles

All the KIR genes present at higher frequencies in CD patients belong to B haplotypes, which contrasts with the more common A haplotype in Caucasian populations. PTEN M-Neg CD patients carried Bx genotypes (i.e. those with at least one B haplotype), including some rare variants (Fig. 3a).

image

Figure 3. (a) Frequency distribution of killer cell immunoglobulin-like receptors (KIR) genotypes in phosphatase and tensin homologue deleted on chromosome 10 (PTEN) mutation-negative (M-Neg) Cowden disease (CD) patients versus Northern Irish Caucasian controls. PTEN M-Neg CD patients display a high incidence of B haplotypes, which contrasts with the more common A haplotypes in Caucasian populations. Bars represent group or population percentages. (b) Individual peripheral blood lymphocyte (PBL) responses to ionomycin (1 µg/ml) for PTEN M-Neg CD patients 1–5 are shown versus control mean. KIR/human leucocyte antigen genotyping data for PTEN M-Neg CD patients 1–5 are shown in Tables 1 and 2. (c) Individual PTEN M-Neg CD PBL IFN-γ secretion responses to Cytostim (20 µl/ml). (d) Individual PTEN M-Neg CD natural killer cell IFN-γ secretion responses to ionomycin (1 µg/ml).

Download figure to PowerPoint

HLA genotyping

HLA-B and -C alleles were assayed. At the B locus, there was a high incidence in the CD kindred of B*27. Other alleles found were B*40, B44 and B07 (Table 2). All patients had C alleles bearing the C2 epitope (K at position 80), whereas only four had the C1 epitope (N at position 80) (Table 2). Specific alleles found included Cw*05, Cw*15, Cw*07 and Cw*02. Interestingly, all the patients lacking Bw4 were heterozygous C1–C2. Owing to the strong linkage of B and C loci, the most probable haplotypes for those alleles are B27 with Cw15 and B40 with Cw02.

Compound KIR/HLA genotypes

To investigate the presence or absence of the KIR cognate ligands, HLA-B and -C, allele distributions for these loci were determined. Inhibitory KIR/HLA associations including KIR3LD1-HLA-Bw4 and KIR2DL3-HLA-C1 were decreased while potentially activating KIR2DS1-HLA-C2 and KIR 3DS1-HLA-Bw4 associations were increased in the PTEN M-Neg CD kindred versus healthy controls. Study of the mutually exclusive Bw4 or Bw6 epitopes revealed three patients with absence of the Bw4 epitope in whom the inhibitory KIR3DL1 was also absent. However, its activating paralogue KIR3DS1 was detected uniformly, so that KIR 3DS1-HLA-Bw4 associations were present in eight PTEN M-Neg CD patients (Tables 1 and 2), which is uncommon in the healthy control population (Table 3) (http://www.allelefrequencies.net).

KIR/HLA compound genotypes varied among PTEN M-Neg CD patients in expression of KIR2DS3, KIR2DS4 as well as HLA-B and HLA-C alleles (Tables 1–3). Patient CD4 expressed >3 KIR activating genes as well as potentially activating KIR2DS1-HLA-C2, KIR 3DS1-HLA-Bw4 compound genotypes, but lacked the inhibitory KIR3DL1 and the KIR2DL3-HLA-C1 association. This patient showed high lymphocyte secretion of IFN-γ after treatment by ionomycin (Fig. 3b) or Cytostim (Fig. 3c) and high NK cell IFN-γ secretion after ionomycin treatment (Fig. 3d).

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References

The present study assayed secreted IFN-γ by cell surface staining and flow cytometry in unfractionated PBLs and isolated NK cell populations in the CD kindred versus healthy controls, in response to both ionomycin as well as the proprietary activating stimulant Cytostim (Miltenyi Biotec). A higher proportion of PTEN M-Neg lymphocytes secreted significantly greater IFN-γ after ionomycin treatment. In a separate experiment effects of Cytostim treatment were broadly similar, inducing high IFN-γ secretion by PTEN M-Neg lymphocytes. Absolute numbers of peripheral blood NK cells were lower in the PTEN M-Neg CD kindred versus healthy controls, possibly as a result of their trafficking to tissues affected by chronic inflammation in the breast or thyroid that is common in CD [1]. Although IFN-γ production by PTEN M-Neg NK cells after ionomycin treatment also appeared higher, differences were not statistically significant.

As IFN-γ production can be enhanced by activating KIRs [13], we assessed KIR/HLA genetics in the PTEN M-Neg kindred versus healthy controls. The PTEN M-Neg patients had higher than expected frequencies of most of the activating KIR genes (KIR2DS1, KIR2DS2, KIR2DS5 and KIR3DS1), together with absence of functional inhibitory KIR3DL1 contained within Bx-haplotypes that are uncommon in Caucasian populations [25]. Moreover, increased frequencies of potentially activating KIR2DS1-HLA C2 with a reduction of inhibitory KIR2DL3-HLA C1 associations were observed, in comparison to the control population. Among the HLA molecules assayed, we found a relatively high incidence of HLA-B*27 (four of 11), HLA-B*07 (four of 11) HLA-B40 (seven of 11) that associate with autoimmunity [37] that characterizes CD model systems [38], as well as component disorders of the CD spectrum including thyroid disease [39] and breast cancer [40]. Functional associations between IFN-γ production and KIR/HLA genetics were explored in individual members of the PTEN M-Neg kindred. Interestingly, the highest lymphocyte IFN-γ secretion after ionomycin or cytostim and the highest NK cell IFN-γ response to ionomycin was observed in CD patient 4, who had extensive KIR activating genes, potentially activating KIR2DS1-HLA-C2, KIR 3DS1-HLA-Bw4 compound genotypes, but lacked the inhibitory KIR2DL3-HLA-C1 association.

In summary, our data show increased IFN-γ responsiveness in association with KIR/HLA activating genotypes in a PTEN M-Neg CD kindred. Genotype–phenotype correlations and definitive linkage associations were precluded in the present study. Furthermore, the overlap in distribution of KIR/HLA markers among kindred members impeded definitive assessment of their individual role in PTEN M-Neg CD risk. While our findings could potentially have wider relevance for the pathobiology of PTEN M-Neg CD or Cowden-like syndromes, where patients have some CD features but do not meet ICC diagnostic criteria [5], further larger-scale studies of unrelated cohorts are warranted.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References

We thank the family members of the study kindred and their clinicians. We also thank Professor Charis Eng (now based at the Cleveland Clinic Main Campus, Mail Code NE5, 9500 Euclid Avenue, Cleveland, OH 44195, USA) for the extensive genetic testing of the PTEN gene and locus in this family. We thank Dr Asensio Gonzalez and Professor D. Middleton, Regional Histocompatibility and Immunogenetics Laboratory, City Hospital, Belfast, for KIR/HLA profiling. This work was supported by grant funding from the HW Baillie Trust and the Northern Ireland Research and Development Office, who are gratefully acknowledged.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References