Blood Journal
Leading the way in experimental and clinical research in hematology

Genomic structure of the PIK3CG gene on chromosome band 7q22 and evaluation as a candidate myeloid tumor suppressor

  1. Christian P. Kratz,
  2. Brooke M. Emerling,
  3. Jeannette Bonifas,
  4. Winfred Wang,
  5. Eric D. Green,
  6. Michelle M. Le Beau, and
  7. Kevin M. Shannon
  1. 1 From the Department of Pediatrics, University of California, San Francisco; Department of Hematology/Oncology, St Jude Children's Research Hospital, Memphis, TN; Genome Technology Branch, National Human Genome Research Institute, Bethesda, MD; and Section of Hematology/Oncology, Department of Medicine, and The Cancer Research Center, University of Chicago, IL.


PIK3CG, which encodes the catalytic subunit p110γ of phosphoinositide 3-OH-kinase-γ (PI3Kγ), has been assigned to chromosome band 7q22, a region that is frequently deleted in myeloid malignancies. PI3Kγ-mutant mice have hematologic defects and are predisposed to colon cancer. On the basis of these data, PIK3CG was evaluated as a candidate myeloid tumor suppressor gene (TSG). PIK3CG was mapped by fluorescence in situ hybridization adjacent and telomeric to a commonly deleted segment defined previously in myeloid leukemias with breakpoints within 7q22. PIK3CG contains 10 exons and spans approximately 37 kilobases of genomic DNA. Forty leukemias with monosomy 7 or a del(7q) were screened for PIK3CG mutations. Two patients had missense variations affecting residue 859 in the N-terminal catalytic domain of the protein. This allele was also detected in unaffected parents and in 1 of 60 control alleles; it probably represents a polymorphism. PIK3CG is unlikely to act as a recessive TSG in myeloid leukemias with monosomy 7.


Phosphoinositide-3-OH kinases (PI3Ks) are ubiquitous lipid kinases that function downstream of cell-surface receptors and in constitutive intracellular membrane and protein-trafficking pathways.1 Hyperactivation of class 1A PI3Ks, which transduce proliferative and antiapoptotic signals from activated tyrosine kinase receptors, is strongly implicated in tumorigenesis in humans and mice. Class 1A PI3Ks consist of a catalytic 110-kd subunit (p110α, β, or δ) and a regulatory 85-kd subunit (p85). PI3Kγ, a class 1B PI3K that links activated G-coupled protein receptors to downstream targets, includes a catalytic subunit (p110γ) that is related to class 1A family members but associates with a novel adapter known as p101. A human complementary DNA (cDNA) encoding p110γ was first isolated from myeloid cells,2 andPI3Kγ-mutant mice show defects in chemotaxis and thymocyte development.3-5 Unexpectedly, Sasaki and coworkers6 found that homozygous PI3Kγ-mutant mice are also strongly predisposed to colon cancer and that expressingPI3Kγ suppressed the growth of established colon cancer cell lines. These investigators mapped PIK3CG, the human homolog of PI3Kγ, to chromosome band 7q22 immediately distal to the RELN and ORC5Lgenes.6

Monosomy 7 and del(7q), or −7/del(7q), are among the most common cytogenetic aberrations identified in myeloid malignancies and occur in a wide spectrum of de novo and secondary cases of acute myeloid leukemia (AML) and myelodysplastic syndrome.7-9 As cytogenetic deletions are a hallmark of tumor suppressor gene (TSG) inactivation,10 it has been hypothesized that chromosome 7 harbors one or more myeloid TSGs. Fluorescence in situ hybridization (FISH) analysis of myeloid malignancies with breakpoints within 7q22 previously defined an approximately 2-megabase (Mb) commonly deleted segment (CDS) bounded by the genetic markers D7S1503 and D7S1841 that contains RELN and ORC5L. 11 12 Studies performed by other researchers also support the importance of 7q22 loss in leukemogenesis.13-16 Together, the hematologic and colon cancer phenotypes found in PI3Kγ-mutant mice and localization of the human homolog within 7q22 led us to evaluatePIK3CG as a candidate TSG in myeloid malignancies with −7/del(7q).

Study design


DNA was extracted from marrow or blood mononuclear cells collected from patients with myeloid malignancies associated with −7/del(7q), as described elsewhere,12 and from the leukocytes of healthy individuals. The experimental protocols involving human subjects were reviewed and independently approved by Committees for the Protection of Human Subjects at the University of Chicago and at the University of California, San Francisco.


The cDNA sequence of PIK3CG (GenBank NM_002649) was used as input data for a blastn search of the National Center for Biotechnology Information database to identify bacterial artificial chromosome (BAC) clones containing PIK3CG.Intron/exon boundaries were assigned by comparing composite cDNA sequences with BAC genomic sequences with the use of theMacvector and AssemblyLIGN programs.

Mutation analysis and sequencing

With the use of genomic DNA as template, exon 1 (approximately 2 kilobases [kb]) was amplified by polymerase chain reaction (PCR) and screened by means of a coupled in vitro transcription/translation (IVTT) procedure to detect mutations resulting in truncated peptides. Exons 2 through 10 were screened by single-strand conformation polymorphism (SSCP) analysis. Descriptions of the SSCP and IVTT procedures employed in our laboratory have been published.17 18 SSCP-PCR products that gave rise to mutations were cloned (TA Cloning Kit, Invitrogen, Carlsbad, CA) and were sequenced by means of the ABI Prism BigDye Terminator Cycle kit (Perkin-Elmer, Norwalk, CT), and the reactions were analyzed with an ABI Prism 310 DNA Sequencer (Perkin-Elmer).

Fluorescence in situ hybridization

Interphase cells were prepared from mitogen-stimulated lymphocytes, and FISH was performed as described previously.11 Labeled BAC probes RG126M09(ORC5L), CTB-152G17 (SRPK2), and GS223D04(PI3KCG), were prepared by nick-translation with the use of Bio-16-dUTP (Enzo Diagnostics, New York, NY) or digoxigenin-11-dUTP (Boehringer Mannheim, Mannheim, Germany). Biotin-labeled probes were detected with fluorescein-conjugated avidin (Vector Laboratories, Burlingame, CA), and digoxigenin-labeled probes were detected with rhodamine-conjugated antidigoxigenin antibodies (Boehringer Mannheim).

PI3KCG expression in normal and leukemia cells with monosomy 7

Quantitative real-time PCR was employed to measurePIK3CG expression in normal and leukemia bone marrows. A Qiagen (Valencia, CA) RNAeasy kit was used to prepare RNA from normal and leukemia marrows, and cDNA was synthesized by means of the SUPERSCRIPT First-Strand Synthesis System for reverse-transcriptase PCR (GIBCO/BRL, Rockville, MD). Primer Express software (PE Applied Biosystems, Foster City, CA) was used to design the following primers for amplifying and probe for TaqMan hybridization: forward (5′-AATTCTCAACTCCCCGAAAGCT-3′); reverse (5′-GGATCGGCACATTTAAACTCAA-3′); probe (5′-FAM-TGGCCTCCAAGAAAAAACCACTATGGC-TAMARA-3′). Glyceraldehyde phosphate dehydrogenase (GAPDH) expression was measured in each sample to ascertain cDNA quality and to establish a reference standard for making comparisons between cases. This gene was selected based on its successful use in previous studies that measured minimal residual disease levels in human leukemias.19 We purchased all reagents from PE Applied Biosystems. The samples were amplified in an ABI Prism 7700 Sequence Detection System at the following thermal cycle parameters: 50°C for 2 minutes, 95°C for 10 minutes, followed by 40 cycles of 95°C for 15 sec and 60 °C for 1 minute.

Results and discussion

The entire PIK3CG coding sequence lies within BAC clone GS223D04 (GenBank AC005018). The cDNA and BAC sequences were aligned, and the GT-AG rule was applied to define intron/exon boundaries. The 3.3-kb PIK3CG cDNA is organized into 10 exons that span approximately 37 kb of genomic DNA (Figure1A). A blastn sequence comparison between BAC GS223D04 and 18 genomic BAC and P1 artificial chromosome clones that we have assigned to the approximately 2-Mb 7q22 CDS did not reveal overlaps, and GS223D04 did not contain any sequence-tagged site markers from this region.12 However, we were unable to conclusively exclude PIK3CG from the CDS using bioinformatic analysis as there were gaps in the region. Cohybridization of BACs containing ORC5L(RG126M09), SRPK2 (CTB-152G17), and PI3KCG(GS223D04) to interphase nuclei using multiple labeling schemes revealed that SRPK2 was located between ORC5L andPI3KCG. A labeling scheme that defines the order of the clones is shown in Figure 1B; it detects SRPK2 (CTB-152G17) with fluorescein and the other 2 BAC clones with rhodamine. Of 200 loci (100 nuclei) analyzed, the order ORC5L-SRPK2-PI3KCG was observed in 102 (51%); the signals were nonlinear (unscorable) in 59 (29.5%), and the order was ORC5L-PI3KCG-SRPK2 in the remaining 39 (19.5%). Subsequent to these FISH studies, data from the Human Gene Project has confirmed that PI3KCG lies approximately 2.2 Mb telomeric of the 7q22 CDS (; accessed July 2001).

Fig. 1.

Genomic structure of PIK3CG, FISH, and mutation analysis.

(A) Simplified diagram illustrating the genomic structure ofPIK3CG and areas within the coding sequence that encode the different p110γ domains. RBD indicates Ras binding domain. The table provides exact information on the exon/intron boundaries within genomic sequence (BAC GS223D04) starting with the first ATG of the open reading frame. (B) BACs RG126M09 (ORC5L) and GS223D04(PI3KCG) were labeled with digoxigenin and detected with rhodamine-conjugated antidigoxigenin antibodies; BAC CTB-152G17(SRPK2) was labeled with biotin and detected with fluorescein-conjugated avidin. In each nucleus, the signal forSRPK2 (green) is located between the signals for the other 2 BACs (red). Nuclei are counterstained with 4,6-diamidino-2-phenylindole-dihydrochloride. (C) Autoradiography showing the SSCP results with the intronic primers flanking exon 6. Arrows indicate normal bands. Leukemias L1 and L2 showed abnormal bands that were inherited in both cases. M indicates mother; F, father). (D) Electropherogram showing the underlying genetic variation ACT>GCT [Thr→Ala] in codon 859 ofPIK3CG.

Because a number of distinct regions of 7q that might harbor myeloid TSGs have been defined, we screened leukemia samples from 40 patients with −7 or a del(7q) for PIK3CG mutations. SSCP with intronic primers flanking exon 6 showed an inherited abnormal SSCP pattern in leukemias L1 and L2 (Figure 1C) that was caused by an ACT>GCT (Thr→Ala) change in codon 859 (Figure 1D). Whereas the normal band was absent in L2, sample L1 displayed both normal and variant bands, owing to the presence of bone marrow cells with and without monosomy 7 as verified by cytogenetic analysis. This exon 6 substitution was also found in 1 of 60 alleles from a control population. Exon 6 encodes a region within the N-lobe of the kinase domain; however, residue 859 is not conserved between PI3K family members.1 As the identical alteration is present in 2 patients, 2 unaffected parents, and 1 normal chromosome, it is likely that this base change in codon 859 represents a single nucleotide polymorphism rather than a pathologic mutation. In addition to this variation in codon 859, one conservative single nucleotide polymorphism, GAC>GAT (Asp) in codon 949, was found in 2 of 40 patients and 3 of 60 normal alleles (data not shown).

Real-time quantitative PCR assay was used to evaluate the possibility that PIK3CG is silenced by epigenetic mechanisms in myeloid leukemia as reported for certain TSGs in other human cancers.20 21 PIK3CG levels in 9 normal bone marrow cells had an average ΔCT of 6.79 relative to GAPDH, with a ΔCT range of 5.9 to 7.9. Fifteen monosomy 7 leukemia samples from which adequate RNA was available for analysis showed similar levels of PIK3CG expression (mean ΔCT, 6.6; range, 5.0-8.7).

Although PI3Kγ-mutant mice are susceptible to colon cancer rather than leukemia, inactivation of TSGs such as p53 orRB1 is associated with a different tumor spectrum in mouse and human.22 23 PIK3CG was an appealing candidate TSG in myeloid leukemia on the basis of its expression in hematopoietic cells and its location within 7q22. However, extensive investigation uncovered neither inactivating mutations ofPIK3CG nor evidence of epigenetic inactivation in human leukemias. Together, these data do not support a major role for biallelic inactivation of PIK3CG in leukemogenesis. The missense change detected in 2 patients was also present in their unaffected parents and in an unrelated healthy individual. This allele is therefore likely to represent a polymorphism, particularly as it neither alters a conserved amino acid nor leads to a truncated protein. Furthermore, Sasaki et al6 surprisingly found that p100γ kinase activity was not required to inhibit the growth of colon cancer cell lines. Our data do not exclude the possibility that heterozygous inactivation of PIK3CG contributes to leukemogenesis, as has recently been reported for the human RUNX1 (AML1)gene.24 25 Haploinsufficiency for the p53 andp27 TSGs have also been implicated in murine tumorigenesis.26 27 However, we have mappedPIK3CG distal to a previously defined CDS. Together, these data and the absence of inactivating mutations in patient specimens support the existence of another gene within 7q22 that functions as a myeloid TSG.


This work was facilitated by a collaboration with the Children's Cancer Group (Study No. B24).


  • Kevin M. Shannon, University of California at San Francisco, 513 Parnassus Ave, HSE 302, San Francisco, CA 94143; e-mail:kevins{at}

  • Supported by National Institutes of Health grant P01 40046 (M.M.L., K.M.S.); the Frank A. Campini Foundation; and a fellowship grant from the Dr Mildred Scheel Stiftung für Krebsforschung (C.P.K.).

  • The publication costs of this article were defrayed in part by page charge payment. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 U.S.C. section 1734.

  • Submitted December 22, 2000.
  • Accepted August 23, 2001.


View Abstract