• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of jclinpathJournal of Clinical PathologyCurrent TOCInstructions for authors
J Clin Pathol. Oct 2006; 59(10): 1059–1065.
Published online Apr 27, 2006. doi:  10.1136/jcp.2005.031161
PMCID: PMC1861748

Comparison of gene‐expression profiles in parallel bone marrow and peripheral blood samples in acute myeloid leukaemia by real‐time polymerase chain reaction

Abstract

Background

Gene signatures (Indicator genes) in bone marrow that provide more precise prognostication in haematological malignancy have been identified by microarray expression studies. It would be beneficial to measure these diagnostic signatures in peripheral blood.

Aims

To determine the degree of correspondence of gene expression for a set of Indicator genes between bone marrow and peripheral blood in acute myeloid leukaemia (AML).

Methods

Parallel bone marrow aspirate and peripheral blood samples were obtained from 19 patients diagnosed with AML and mononuclear cells isolated from both sample types. mRNA was globally amplified by polyadenylated real‐time polymerase chain reaction (polyA RT‐PCR); the expression of 15 AML Indicator genes, identified from previous microarray studies, was measured by RT‐PCR. All values were normalised to the mean expression of three housekeeping genes (IF2‐β, GAP and RbS9) and were statistically compared using SPSS software.

Results

No significant difference in expression between bone marrow and peripheral blood was observed for 10 of the genes (leptin receptor, CD33, adipsin, proteoglycan 1, MB‐1, cyclin D3, hSNF2b, proteasome iota, HkrT‐1 and E2A), indicating its possible use in monitoring disease activity in peripheral blood samples, whereas c‐myb, HOXA9, LYN, cystatin c and LTC4s showed significantly different expression between bone marrow and peripheral blood samples.

Conclusion

These results indicate a possible use for the method in monitoring AML in peripheral blood by RT‐PCR measurement of Indicator genes. In addition, the initial use of polyA PCR facilitates translation to very small clinical samples, including fractionated cell populations, of particular importance for monitoring haematological malignancy.

Diagnosis and monitoring of the acute leukaemias, acute myeloid leukaemia (AML) and acute lymphoblastic leukaemia (ALL), is currently achieved at the level of cell morphology, protein expression and cytogenetics.1 Recent studies have shown the ability to further stratify disease on the basis of expression of gene signatures, Indicator genes, identified by microarray profiling.2 These studies hold the promise of more specific diagnosis, prognostication and the development of tailored treatment,3 and a practical method is now required for the measurement of these gene sets in routine clinical practice.

However, use of cDNA arrays in routine clinical practice is limited by cost and the need for relatively large amounts of RNA. We have used a global amplification approach known as polyadenylated polymerase chain reaction (polyA PCR) to overcome these limitations. PolyA PCR coordinately amplifies cDNA copies of all polyadenylated mRNAs, thereby generating a PCR product (polyA cDNA) whose composition reflects the relative abundance of all expressed genes in the starting sample, and which can be applied to very small samples, including single cells.4 Real‐time polymerase chain reaction (RT‐PCR) measurement using gene‐specific primers and probes of the expression levels of specific Indicator genes then allows gene signatures to be detected in the polyA cDNA, thereby enabling expression profiling of very small amounts of starting material.5 We have tested use of the method by measuring levels of gene expression in bone marrow from patients with acute leukaemia for 17 Indicator genes selected from a previous microarray study by Golub et al,6 most (n = 15) of which showed similar expression in AML and ALL to that reported by Golub et al,6 suggesting diagnostic utility.7 The aim of our study was to determine the degree of correspondence of gene expression between bone marrow and peripheral blood for the same set of Indicator genes as a first step towards their use for diagnosis using peripheral blood rather than bone marrow samples.

Materials and methods

Sample acquisition

Parallel bone marrow aspirate and peripheral blood samples, both collected from each patient at the same time, were obtained, with informed consent, from 19 patients with AML, after treatment (10 men, 9 women; age range 20–72 years, mean age 45 years). Table 11 gives the clinical details for each patient, including the blast count in the parallel bone marrow and peripheral blood samples in each case; full details were not available for one patient. Table 11 also details the age and cytogenetic abnormality for each patient at the time of primary diagnosis of AML and at the time of sampling of parallel bone marrow aspirate and peripheral blood samples. For all but one of the patients, samples were collected during routine monitoring while the patients were in remission. All bone marrow aspirates were collected in Hank's buffered saline solution with heparin, whereas peripheral blood samples were collected in tubes containing standard ethylenediaminetetraacetic acid. Mononuclear cells were isolated immediately from both samples by density‐gradient centrifugation and used for RNA extraction.

Table thumbnail
Table 1 Clinical details of patients and samples

Global amplification of polyadenylated mRNAs (polyA RT‐PCR)

Total RNA was extracted from all the samples, using an RNeasy mini kit (Qiagen, Valencia, California, USA). Global amplification of cDNA corresponding to all expressed genes (polyA PCR) was carried out as previously reported.4,5,8

Taqman RT‐quantitative PCR

Golub et al6 identified 50 Indicator genes discriminatory between AML and ALL, from which 15 Indicator genes had been chosen for measurement by RT‐PCR in bone marrow aspirate samples from patients with AML and ALL in our previous study.7 These genes included genes highly expressed in AML—cystatin c, leptin, CD33, HOXA9, adipsin, proteoglycan 1, LTC4s and LYN, and genes highly expressed in ALL—c‐myb, MB‐1, cyclin D3, hSNF2, proteasome iota, HkrT‐1 and E2A. Taqman PCR primers and probes were designed for each gene, using Primer Express Software (Perkin Elmer Applied Biosystems, Foster City, California, USA), and are listed in table 22.. For each gene, Taqman PCR was applied to 1 ng of polyA cDNA from each sample and to 10 μl of serially diluted human genomic standards, using a Taqman Gold kit.5 All samples were analysed with an ABI Prism 7700 sequence detection system (Perkin Elmer Applied Biosystems). The copy number of each gene was determined by reference, after normalisation to the mean of the three housekeeping genes (Mhouse; below), of the RT‐PCR expression level to the human genomic DNA standards.5

Table thumbnail
Table 2 List of TaqMan primers and probes

Normalisation

The expression levels of three housekeeping genes (IF2‐β, GAP and human ribosomal protein S9 (RbS9) mRNAs) were measured by RT‐PCR in each sample. Copy numbers obtained for the Mhouse (IF2‐β, GAP and RbS9) in each sample were divided by the highest Mhouse in all samples, resulting in a normalisation correction factor. After real‐time PCR amplification and quantification of the selected genes, this factor was used for normalisation of expression levels of each of the 15 genes measured.

Statistical analysis

The data were not normally distributed and non‐parametric tests were used. Statistical analysis of the expression levels of the 15 genes in the two groups (bone marrow aspirate v peripheral blood) was carried out using the Mann–Whitney U test with p[less-than-or-eq, slant]0.05 for significance. All the tests were carried out using SPSS software. As the expression levels for all the genes have a wide range, the natural logarithm was used to plot all values because it allows a wide range of values to be visually compared. The ranking of all the expression levels of Indicator genes between the paired peripheral blood and bone marrow aspirate samples was analysed and presented as the mean rank statistical difference.

Cluster analysis

Unsupervised cluster analysis of the normalised gene expression values for each sample was carried out using Cluster and TreeView from the Eisen laboratory (http://rana.lbl.gov/EisenSoftware.htm), with the protocols detailed with the software. Clustering was carried out on the gene expression values from all 38 samples for all genes, and separately for those showing significant difference in expression between bone marrow and peripheral blood samples with the Mann–Whitney U test.

Results

PolyA real‐time polymerase chain reaction

All 38 samples of total RNA prepared from the 19 paired samples of human bone marrow and peripheral blood were subjected to polyA RT‐PCR, yielding polyA cDNA in the expected size range of 100–700 bp.

Comparative gene expression pattern of housekeeping genes in parallel bone marrow and peripheral blood samples

Figure 11 shows the mean expression levels for the three housekeeping genes (IF2‐β, GAP and RbS9) in the bone marrow aspirate and peripheral blood samples, with the Mhouse. All three genes were expressed in both samples and all showed higher mean expression in the peripheral blood than in the bone marrow, particularly RbS9. The mean expression levels of GAP and RbS9 were much higher than that of IF2‐β (fig 11).

figure cp31161.f1
Figure 1 Comparative gene expression pattern of IF2‐β, GAP and Rbs9 housekeeping genes in parallel bone marrow and peripheral blood samples. The mean expression levels of the three housekeeping genes, measured by RT‐PCR ...

Comparison of Indicator gene expression levels in parallel bone marrow and peripheral blood samples

C‐myb, HOXA9, LYN, cystatin c and LTC4s showed a significant difference in expression in the bone marrow and peripheral blood samples, whereas the other genes did not. The expression values for the significantly different genes are shown in box plots (fig 22).). The levels of expression of HOXA9 and LTC4s in the bone marrow samples were significantly higher in the bone marrow than in the peripheral blood, whereas by contrast, the levels of expression of cystatin c and LYN were significantly lower in the bone marrow than in the peripheral blood.

figure cp31161.f2
Figure 2 Expression levels of genes with significant difference between bone marrow and peripheral blood. Box plots show the values for the genes for which there was a significant difference in the level of expression between the bone marrow and ...

Figure 33 shows the mean ranks for the Indicator genes between the two samples, which indicate divergence of those genes with significant difference between the samples and convergence of the remainder.

figure cp31161.f3
Figure 3 Mean ranks for the Indicator genes in acute myeloid leukaemia (AML) and acute lymphoblastic leukaemia (ALL) in each of the bone marrow and peripheral blood samples. Mean ranks (y axes) for each gene, calculated using the Mann—Whitney ...

Cluster analysis

Cluster analysis carried out in the space of all genes showed no clear clustering of the samples, although the housekeeping genes (IF2‐β, GAP and RbS9 mRNAs) clustered together (fig 4A4A).). Conversely, unsupervised clustering in the space of those genes showing significant difference in expression between bone marrow and peripheral blood samples (c‐myb, HOXA9, LYN, cystatin c and LTC4s) showed clear clustering of the bone marrow and peripheral blood samples, confirming that these genes differ between the two samples (fig 4B4B).). Clustering in the space of those genes with no significant difference in expression between bone marrow and peripheral blood showed no clear grouping of the two samples (fig 4C4C).

figure cp31161.f4
Figure 4 Unsupervised cluster analysis for (A) all genes, (B) genes showing either the presence or the absence of significant differences in expression between bone marrow and peripheral blood and (C) genes with significant differences in expression ...

Discussion

Several microarray studies have identified gene signatures for haematological malignancies in general, and for acute leukaemia in particular.2,3 These promise more specific diagnosis, prognosis and treatment, with expected improvements in patient survival,3 and a practical method for measurement of these gene signatures in routine clinical practice is urgently required.

We have used a global amplification approach known as polyA PCR to overcome these limitations. PolyA PCR coordinately amplifies cDNA copies of all polyadenylated mRNAs, thereby generating a PCR product (polyA cDNA), the composition of which reflects the relative abundance of all genes expressed in the starting sample, and which can be applied to very small samples, including single cells.4 The expression profiles of specific Indicator genes can then be measured by RT‐PCR using gene‐specific primers and probes, thereby enabling expression profiling of very small amounts of starting material.5 We have tested the use of the method in bone marrow samples from AML and ALL, showing its ability to distinguish between these two groups using Indicator genes identified from independent microarray studies.6,7 Specifically, for 17 Indicator genes showing differential expression between AML and ALL by microarray analysis,6 all but two genes measured by RT‐PCR showed expression in AML and ALL similar to that in the microarray study.7 Following on from this work, the aim of our study was to determine the degree of correspondence of gene expression for the same set of Indicator genes between bone marrow and peripheral blood samples as a first step towards their use for diagnosis using peripheral blood either together with, or instead of, bone marrow samples.

To do this we measured, by RT‐PCR, the expression levels of 15 Indicator genes for AML in 19 parallel bone marrow and peripheral blood samples. Each of these sample pairs was harvested from patients treated for AML, all but one of whom was in remission at the time of sampling. The 15 genes measured were selected from genes discriminatory between AML and ALL.6 Although all the paired bone marrow and peripheral blood samples analysed in our study were from patients with AML, the seven Indicator genes for ALL were included, as patients were recruited to the study sequentially and it was not possible to predict in advance the number of patients presenting with either AML or ALL.

Comparison of the median levels of expression for each gene between the two sample types showed significant differences in the level of expression for 5 of the 15 genes measured—namely, c‐myb, cystatin c, HOXA9, LTC4s and LYN, whereas the remaining 10 genes (adipsin, CD33, proteoglycan, proteosome iota, leptin, cyclin D3, E2A, SNF2, HkrT‐1 and MB‐1) showed no significant difference in expression levels between the two sample types. This indicates that the expression level for these 10 genes in the peripheral blood reflects their expression in the bone marrow, suggesting therefore that the gene‐expression profile in the bone marrow can be judged by peripheral blood analysis. Most samples used were from patients in remission. Notwithstanding this, there remained a good correlation between the expression levels for these 10 genes between the two sample types. Such a correlation was therefore not a feature merely of high blast counts, and suggests that the level of these 10 genes in the peripheral blood reflects that in the bone marrow even at low blast counts. This may facilitate monitoring of acute myeloid leukaemia, particularly if the genes measured are associated with other known molecular markers of minimal residual disease.

Cluster analysis confirmed the results of the Mann–Whitney U analysis by an independent, unsupervised approach. Specifically, those genes showing significant difference in expression between the two sample types were able to separate the two types by unsupervised clustering, whereas no distinction between the two types was seen by clustering in the space of those genes showing no significant difference between bone marrow and peripheral blood samples. Conversely, the housekeeping genes clustered together, showing equivalence of their expression across all samples, thereby validating their use.

Of interest is the fact that of the five genes showing a difference between the two sample types (cyclin D3, proteoglycan, LYN, c‐myb and HOXA9), only one, c‐myb, is ALL specific, whereas the remaining four were all drawn from the list of genes up regulated in AML compared with ALL. This is of relevance considering that all the samples analysed were from patients with AML, in whom it may be hypothesised that genes predictive of ALL would probably not be altered from normal, either in the bone marrow or in the peripheral blood. Conversely, for the other four genes, difference in expression level between the peripheral blood and bone marrow may be due to the particular involvement of these genes in the pathobiology of AML, resulting in their differential expression between bone marrow and circulating leukaemic cells.

The results show the utility of the polyA PCR method for generation of polyA cDNA from bone marrow and peripheral blood. This aspect of the method renders it applicable to very small clinical samples, by virtue of the amplification by polyA PCR. The Indicator genes measured showed similarity of expression between bone marrow and peripheral blood for most (10 of 15) of the genes, indicating the possible use of the method for monitoring of acute myeloid leukaemia in the peripheral blood by RT‐PCR measurement of AML Indicator genes.

Abbreviations

AML - acute myeloid leukaemia

ALL - acute lymphoblastic leukaemia

Mhouse - mean of the three housekeeping genes

polyA PCR - polyadenylated polymerase chain reaction

RbS9 - ribosomal protein S9

RT‐PCR - real‐time polymerase chain reaction

Footnotes

Funding: This work was supported by a National Heath Service Research & Development grant.

References

1. Vardiman J W, Harris N L, Brunning R D. The World Health Organization (WHO) classification of the myeloid neoplasms. Blood 2002. 1002292–2302.2302 [PubMed]
2. Ebert B L, Golub T R. Genomic approaches to hematologic malignancies. Blood 2004. 104923–932.932 [PubMed]
3. Grimwade D, Haferlach T. Gene‐expression profiling in acute myeloid leukemia. N Engl J Med 2004. 3501676–1678.1678 [PubMed]
4. Brady G, Barbara M, Iscove N N. Representative in vitro cDNA amplification from individual haemopoietic cells and colonies. Methods Mol Cell Biol 1990. 217–22.22
5. Al‐Taher A, Bashein A, Nolan T. et al Global cDNA amplification combined with real‐time RT‐PCR: accurate quantification of multiple human potassium channel genes at the single cell level. Yeast 2000. 17201–210.210 [PMC free article] [PubMed]
6. Golub T R, Slonim D K, Tamayo P. et al Molecular classification of cancer: class discovery and class prediction by gene expression monitoring. Science 1999. 286531–537.537 [PubMed]
7. Sakhinia E, Faranghpour M, Liu Yin J. et al Routine expression profiling of microarray gene signatures in acute leukaemia by real‐time PCR of human bone marrow. Br J Haematol 2005. 130233–248.248 [PubMed]
8. Byers R, Roebuck J, Sakhinia E. et al PolyA PCR amplification of cDNA from RNA extracted from formalin‐fixed paraffin‐embedded tissue. Diagn Mol Pathol 2004. 3144–150.150 [PubMed]

Articles from Journal of Clinical Pathology are provided here courtesy of BMJ Group
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

  • Gene (nucleotide)
    Gene (nucleotide)
    Records in Gene identified from shared sequence links
  • MedGen
    MedGen
    Related information in MedGen
  • Nucleotide
    Nucleotide
    Published Nucleotide sequences
  • PubMed
    PubMed
    PubMed citations for these articles
  • Substance
    Substance
    PubChem Substance links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...