Evidence for the involvement of a hematopoietic progenitor cell in systemic mastocytosis from single-cell analysis of mutations in the c-kit gene

Mastocytosis is a rare disorder characterized by a pathologic increase in mast cells in tissues including skin, bone marrow, liver, spleen, and lymph nodes. Activating mutations in c-kit, the gene encoding the receptor for Kit ligand, also known as stem cell factor, may be detected in the bone marrow, peripheral blood, and lesional skin in patients with mastocytosis.1,2 These mutations most commonly involve codon 816 of the intracellular tyrosine kinase domain and cause ligand-independent autophosphorylation of the receptor. An additional mutation in codon 560 within the juxtamembrane domain of Kit, yielding a substitution of glycine for valine, has been detected in the human mast cell leukemia line HMC-1³ and, in one study, in some patients with mastocytosis.4 

Recently, messenger RNA (mRNA) encoding the Asp816Val c-kitmutation has been detected in non–mast cell hematopoietic lineages in the peripheral blood and bone marrow in patients with indolent mastocytosis and in those with an associated hematologic disorder.5 This observation suggests that in mastocytosis, the mutation in c-kit occurs in a pluripotential hematopoietic progenitor cell. Lineage relationship and the frequency of the cells derived from the progenitor with the mutated c-kit, however, could not be established because the observations relied on the analysis of bulk mRNA.

The current study was therefore carried out to determine the lineage relationship of cells derived from the hematopoietic progenitor with the c-kit mutation, to assess whether the cells derived from the mutated clone coexist with those carrying the wild-type gene, and to document the involvement of a hematopoietic progenitor cell in disease pathogenesis. To achieve these goals, we examined the occurrence of the Asp816Val and the Val560Gly c-kitmutations in the genomic DNA of individual peripheral blood T cells, B cells, and monocytes in 5 patients with indolent systemic mastocytosis using single-cell mutational analysis. The Asp816Val c-kitmutation was detected in varying frequencies in the genomic DNA of B cells and monocytes in 3 patients with extensive mastocytosis. These results demonstrate that mastocytosis is a clonal disorder of a pluripotential hematopoietic progenitor cell and that the affected clone shows variable expansion in these cell lineages in the peripheral blood of patients with systemic mastocytosis.

Five patients with adult-onset indolent systemic mastocytosis were studied after informed consent according to the Declaration of Helsinki was obtained on a protocol approved by the institutional review board of the National Institute of Allergy and Infectious Diseases (Table 1). Patients had elevated plasma levels of surrogate disease markers, tryptase6 and soluble CD117.7 Three patients (no. 3, 4, and 5; Table 1) had hypercellular bone marrows (> 70% cellularity with myeloid hyperplasia but no overt clonal hematologic non–mast cell lineage disorder) in addition to markedly elevated plasma levels of surrogate markers. The clinical and laboratory features of these patients met the criteria to be classified in the “smoldering mastocytosis” subcategory of indolent systemic mastocytosis.8 

Peripheral blood mononuclear cells were isolated using Ficoll-Hypaque density gradient centrifugation. CD3⁺ T cells, CD19⁺ B cells, and CD14⁺ monocytes were stained with lineage-specific monoclonal antibodies (BD Pharmingen, San Diego, CA). Each cell lineage was first sorted in bulk using the FACStar flow cytometer (Becton Dickinson, San Diego, CA). After this procedure, single cells were sorted by flow cytometry into individual wells of 96-well polymerase chain reaction (PCR) plates (USA Scientific, Bethesda, MD) containing 20 μL 1 times PCR buffer (Promega, Madison, WI) and proteinase-K (0.2 mg/mL; Sigma, St Louis, MO). The purity of each sorted population was more than 99% as determined by flow cytometry.

Mast cells were isolated from a bone marrow aspirate from patient 3 and bone marrow biopsy tissue from patient 4. Bone marrow aspirate mononuclear cells from patient 3 were incubated with CD117-phycoerythrin (PE) and CD25–fluorescein isothiocyanate (FITC) at 4°C for 30 minutes, and mast cells were sorted into individual wells as CD117⁺, CD25⁺ cells with high side-scatter. Bone marrow biopsy tissue from patient 4 was incubated in phosphate-buffered saline (PBS) containing 2 mM EDTA and 0.1% bovine serum albumin (BSA) for 2 hours. It was then disaggregated mechanically by cutting the tissue into small pieces and passing it through a 19-gauge needle several times followed by filtering through a cell strainer. The cell suspension prepared from the bone marrow biopsy tissue was incubated with an anti-CD25 antibody conjugated to PE (BD Pharmingen) at 4°C for 30 minutes, washed once with PBS supplemented with 0.1% BSA, and incubated with anti-PE paramagnetic microbeads (Miltenyi Biotec, Auburn, CA). Cells expressing surface CD25 were then isolated according to manufacturer's instructions. Cytospin preparations obtained from CD25⁺ cells and stained with Wright-Giemsa stain following this procedure showed mast cells and lymphoid cells, as expected. Mast cells were then distinguished from lymphoid cells based on their characteristic high side-scatter pattern and were sorted into the single wells of a 96-well PCR plate with a FACStar flow cytometer.

Cells sorted into 96-well PCR plates were incubated for 1 hour at 55°C and 10 minutes at 95°C. Genomic DNA was then preamplified using random 15-mer primers and 60 rounds of amplification withTaq polymerase, as described.9 The preamplified genomic DNA was then subjected to nested PCR (Table2). Both juxtamembrane and tyrosine kinase domains of c-kit were amplified, sequenced, and analyzed for the presence of Gly560Val and Val816Asp mutations, respectively. The first region (c-kit–R1) containing codon 560 spans exon 10, intron 10, and exon 11. The second region (c-kit–R2) spans exon 17 and intron 17 and contains codon 816.

To assess the presence of mutations, PCR products were purified with QIAquick PCR purification kit (Qiagen, Hilden, Germany) and directly sequenced using an ABI Prism 377 automated DNA sequencer (Applied Biosystems, Foster City, CA) and a Big Dye Terminator DNA sequencing kit (Applied Biosystems, Foster City, CA). The sequences were aligned with the GenBank sequence of human c-kit(accession number, U63834) using the software Lasergene (DNASTAR, Madison, WI).

The PCR error rate in this study was calculated by analyzing 75 182 base pairs (bp) of c-kit sequence obtained from 257 individual cells (79 B cells, 86 monocytes, 92 mast cells) from patient 4. A total of 6 nucleotide errors were detected, indicating that the PCR error rate was 8 × 10⁻⁵/nucleotide.

Immunoglobulin gene V_H/(D)/J_H rearrangements from individual B cells were amplified and sequenced as described.9 

To determine the frequency of the 81517C>T polymorphism of c-kit, bulk genomic DNA was isolated from the peripheral blood mononuclear cells of 22 patients with mastocytosis and 69 healthy donors. Genomic DNA was then subjected to nested PCR using c-kit R2 primers and the conditions shown in Table 2. PCR products were purified, sequenced, and analyzed as described above. A peak consisting of C and T nucleotides in equal heights were observed in individuals carrying the polymorphism as a heterozygous allele. A single peak of either C or T was observed in patients carrying homozygous wild-type or polymorphic alleles.

Initially, a 292-bp tyrosine kinase fragment spanning exon 17 and intron 17 of the c-kit genomic DNA was examined in single-sorted T cells, B cells, monocytes, and mast cells from patients with mastocytosis. Of 1824 cells sorted, 1292 (71%) yielded sequences suitable for analysis. In peripheral blood of patients 3, 4, and 5, the Asp816Val c-kit mutation was detectable in varying frequencies ranging from 3.8% to 89.5% of the CD14⁺monocytes and 10.3% to 55.7% of the CD19⁺ B cells (Table3). The frequency of the monocytes and B cells carrying the mutation was lower in patient 3, who also had a lower serum tryptase level when compared with patients 4 and 5. The c-kit mutation was either absent or detectable in low frequency (in 1 of 60 cells from patient 5) in T cells. In addition, the mutation was detectable in 29 of 110 (26.4%) and 158 of 187 (84.5%) sequences obtained from bone marrow mast cells of patients 3 and 4, respectively.

The single-cell analysis made it possible to amplify and sequence immunoglobulin heavy-chain genes amplified from the same individual B cells carrying the Asp816Val mutations from patient 4. The sequences obtained (n = 23) were polyclonal with no evidence of monoclonal expansion (data not shown). These results indicate that the Asp816Val mutation in c-kit occurred before V/(D)/J recombination, prior to the pro–B-cell stage of B-cell differentiation.

The Asp816Val mutation was not detectable in any of the cell lineages examined in patients 1 and 2. Patients 3, 4, and 5, who had mutations detectable in multiple cell lineages, had disease that was classified as systemic smoldering mastocytosis as characterized by a high degree of mast cell infiltration of the bone marrow, myeloid hyperplasia, and higher plasma levels of tryptase.8 

Because the HMC-1 cell line derived from a patient with mast cell leukemia contains a codon 560 c-kit mutation in addition to the Asp816Val mutation3 and because there is one report of detection of this mutation in patients with mastocytosis,4 we next examined the juxtamembrane codon 560 in exon 10 for mutations in the same DNA samples. We evaluated a total of 468 sequences from patients 1 through 4 including 95 sequences from bone marrow mast cells of patient 4. No mutations were detected in codon 560 in any of the sequences examined.

Analysis of the sequencing results of the PCR products revealed the presence of 2 c-kit polymorphisms. The first of these polymorphisms resulted in a Met→Leu substitution (A75544C) in one patient (no. 3). The frequency of this polymorphism has been reported to be similar in patients with mastocytosis and healthy subjects.10 

In addition, a novel polymorphism (81517C>T) in intron 17 of the c-kit gene was identified. The 81517C>T c-kit polymorphism was detected in 857 of 1054 (81.3%) amplifications from 4 patients (patients 2-5) bearing this polymorphism in a heterozygous state. Because of the apparent high frequency of this polymorphism in our study population, we determined whether this polymorphism was present in increased frequency in patients with mastocytosis as compared to healthy individuals. We thus analyzed genomic DNA from 22 patients with mastocytosis and 69 healthy control subjects. The allelic frequency of this polymorphism was found to be similar in patients with mastocytosis compared to the healthy population (36% versus 31%, respectively). We therefore concluded that 81517C>T c-kit polymorphism is not significantly associated with mastocytosis.

The single-cell technique used in this study made it possible to determine whether the mutation and the polymorphism occurred on the same or different alleles. This was analyzed in patient 4, who had a high frequency of mast cells with the Asp816Val mutation (Figure1). In many mast cells (149 of 187, 80%) both c-kit alleles were amplified and the relation between the mutation and the polymorphism could not be determined (Figure1B,C). However, in other cells (38 of 187, 20%), only one allele was amplified (Figure 1A,D). This made it possible to determine whether the mutation and the polymorphism occurred on the same allele. Notably, the mutation and the polymorphism never (0 of 38) occurred in the same sequence. Thus, 12 sequences contained the mutation but not the polymorphism (Figure 1A), whereas 26 sequences had the polymorphism but not the mutation (Figure 1D). These data strongly imply that the mutation and the polymorphism were found exclusively on different alleles.

The single-cell genomic DNA analysis presented in this study is the first clear documentation of the involvement of a multipotential hematopoietic progenitor cell in systemic mastocytosis. As a result of a mutation in a common hematopoietic progenitor, the mutation could be detected in mast cells, monocytes, and B cells in variable frequencies in 3 patients with extensive disease (Table 1).

The cell lineages found to have the Asp816Val mutation in this study correlate with the reported occurrence of this mutation in bulk mRNA samples obtained from these hematopoietic populations in patients with mastocytosis.5 Thus, detection of the mutation in peripheral blood B cells and monocytes and bone marrow mast cells may be explained by the mutational targeting of c-kit in a progenitor cell with the restricted capacity to differentiate into these lineages. Alternatively, the mutation may occur in a more primitive precursor cell, but it may provide a preferential growth advantage only to B cells, monocytes, and mast cells, although the mutation does not result in accumulation of excessive numbers of B cells and monocytes in the peripheral blood. It is noteworthy, however, that patients with the c-kit mutation detectable in B cells and monocytes did have hepatosplenomegaly and hypercellular bone marrow, indicating that they may have had an increased total body burden of hematopoietic cells. It should also be noted that the c-kit mutation was found in one T cell. Whether this represents a PCR error or indicates that the c-kitmutation also occurs in T cells at a low frequency cannot be determined. However, the data do indicate that the frequency of the Asp816Val mutation in c-kit is much lower in T cells than in B cells and monocytes.

An analogous lineage involvement pattern has been reported in chronic myeloid leukemia (CML) and in myelodysplastic syndromes (MDSs). Thus, combined use of fluorescence-activated cell sorting and fluorescence in situ hybridization revealed the presence of the Philadelphia chromosome in myeloid cells, most B cells, and bone marrow T-cell progenitors, but not in mature peripheral T cells in the chronic phase of CML.11 Similar observations were reported in patients with MDS.12,13 These findings are consistent with the conclusion that systemic mastocytosis, similar to CML and MDS, is a clonal disorder of a hematopoietic progenitor cell with a capacity to give rise to B cells and myeloid cells preferentially. The demonstration of a polyclonal rearrangement pattern for the immunoglobulin heavy chain in B cells carrying the mutation is also consistent with this conclusion.

Our data strongly support the concept that systemic mastocytosis, at least the smoldering subvariant, is closely related to the classical group of chronic myeloproliferative diseases. These diseases are characterized by the clonal involvement of a hematopoietic progenitor cell leading to excessive numbers of mature cells belonging to one or more of the myeloid lineages. The other common features shared by systemic smoldering mastocytosis and chronic myeloproliferative diseases include the frequent presence of hepatosplenomegaly, increased bone marrow cellularity, and the activation of tyrosine kinase signal transduction pathways.

All 3 of the patients who had demonstrable multilineage involvement in this study met the criteria to be included in the recently established “indolent smoldering mastocytosis” category.14-16 This category defines patients with extensive mast cell burden as evidenced by more than 30% infiltration of the bone marrow cavity by mast cells, elevated serum tryptase levels of more than 200 ng/mL, and hepatosplenomegaly. The expansion and the relative frequency of cells clonally derived from the hematopoietic progenitor carrying the c-kit mutation may thus correlate with the extent of the disease. The observation that the frequency of B cells and monocytes carrying the c-kit mutation was higher in patients with higher tryptase levels supports this explanation. A modest growth advantage conveyed by the c-kit mutation could account for the slowly developing expansion of affected hematopoietic lineages.

Our results did not identify the juxtamembrane Val560Gly c-kit mutation in cells carrying the Asp816Val c-kit mutation including lesional bone marrow mast cells. Consistent with this observation, juxtamembrane c-kitmutations generally are described to be associated with gastrointestinal stromal tumors but not mastocytosis.17 

We have identified a novel polymorphism (81517C>T) in intron 17 of c-kit. This polymorphism was present in high frequency in both patients with mastocytosis and in healthy individuals. The occurrence of this polymorphism in a heterozygous state also served as an internal control for the sensitivity of the single cell PCR analysis. This nucleotide change was detectable in 857 of 1054 (81.3%) single-cell PCR products in patients carrying the polymorphism as opposed to the expected frequency of 100% as a germline polymorphism. This is likely to be a result of differential amplification of the heterozygous alleles during the initial nonspecific preamplification and the subsequent nested PCR steps. If the frequency data for c-kit Asp816Val is normalized according to an allelic dropout rate of 18.7%, 100% of the mast cells and monocytes and 68.5% of B cells analyzed in patient 4 can be estimated to carry the mutation.

Although the frequency of the 81517C>T polymorphism was not greater in patients with mastocytosis than in healthy blood donors, the amplification of c-kit genes from individual cells made it possible to demonstrate a potentially unique feature of the role of this polymorphism in mastocytosis. The analysis of c-kitgene in mast cells of patient 4 was particularly informative because nearly all (84.5%) of these contained the Asp816Val mutation. The single-cell PCR technique usually (80%) amplifies both alleles from each cell. Most of the mast cells were clearly heterozygous for the mutation. It was notable, however, that examination of the height of the nucleotide peaks generated by the automated sequencer suggested that there was an inverse correlation between the height of the mutation peak and that of the polymorphism (Figure 1A,B). This suggested that the mutation and the polymorphism resided on different alleles. More compelling was the analysis of the 20% of the cells in which only one allele was amplified. In these cells, there was a complete discordance between the detection of the mutation and the polymorphism, indicating that they occurred on separate alleles. Analysis of the pattern of mutation and polymorphism in patients 3 and 5 revealed a similar pattern. Thus, in 53 mast cells from patient 3 and 75 monocytes from patient 5, the polymorphism and the mutation were never in the same sequence. This clearly documents the precision of the single-cell PCR technique, but also suggests that the development of mastocytosis may require 2 lesions of the c-kit gene. The activating Asp816Val mutation may be required in one allele along with the 81517C>T polymorphism that may serve to alter the expression or function of the other c-kit allele. Analysis of additional patients will be necessary to confirm or reject this hypothesis.

In conclusion, these data provide support for inclusion of systemic mastocytosis in the group of hematopoietic clonal disorders and offer insight into the nature of the progenitor cell affected by the c-kit mutation.