Whole blood vs PBMC: compartmental differences in gene expression profiling exemplified in asthma
Blood has proven to be a useful resource for molecular analysis in numerous biomedical studies, with peripheral blood mononuclear cells (PBMCs) and whole blood being the major specimen types. However, comparative analyses between these two major compartments (PBMCs and whole blood) are few and far between. In this study, we compared gene expression profiles of PBMCs and whole blood samples obtained from research subjects with or without mild allergic asthma.
Whole blood (PAXgene) and PBMC samples were obtained from 5 mild allergic asthmatics and 5 healthy controls. RNA from both sample types was measured for expression of 730 immune-related genes using the NanoString nCounter platform.
We identified 64 uniquely expressed transcripts in whole blood that reflected a variety of innate, humoral, and adaptive immune processes, and 13 uniquely expressed transcripts in PBMCs which were representative of T-cell and monocyte-mediated processes. Furthermore, analysis of mild allergic asthmatics versus non-asthmatics revealed 47 differentially expressed transcripts in whole blood compared to 1 differentially expressed transcript in PBMCs (FDR < 0.25). Finally, through simultaneous measurement of PBMC proteins on the nCounter assay, we identified CD28 and OX40 (TNFRSF4), both of which are critical co-stimulatory molecules during T-cell activation, as significantly upregulated in asthmatics.
Whole blood RNA preserved in PAXgene tubes is excellent for producing gene expression data with minimal variability and good sensitivity, suggesting its utility in multi-centre studies requiring measurement of blood gene expression.
KeywordsWhole blood PBMC Gene expression Asthma
differentially expressed gene
false discovery rate
forced expiratory volume, 1 s
peripheral blood mononuclear cell
provocative concentration for 20% fall in FEV1
polymerase chain reaction
quantitative polymerase chain reaction
standard operating procedure
Unbiased BIOmarkers in PREDiction of respiratory disease outcomes
Gene expression profiling is a fundamental feature of systems biology studies, which are ever-increasing in use to unravel the complexity of biological systems and their disease states. In particular, measurement of gene expression (‘transcriptomics’) in blood is an effective tool for the discovery of biomarkers of disease, and remains a key component of studies requiring large numbers of samples due to ease of collection and storage. Although the advent of single-cell RNA sequencing has made it possible to measure gene expression of specific blood cell types, there remain financial and technical barriers to its widespread use. Hence, the most common methods of blood RNA profiling either investigate RNA extracted from all blood cell types (i.e. whole blood) or only peripheral blood mononuclear cells (PBMCs). While PBMCs (lymphocytes and monocytes) play a significant role in the immune system, they remain a subset of all immune cells and do not include other cell types such as eosinophils and neutrophils. Thus, consideration of the compartment being investigated is required when carrying out blood gene expression studies, particularly with respect to the disease or other condition of interest.
Asthma is a chronic disease of the lungs that is characterized by frequent dyspnea, wheezing, coughing, and chest tightness due to inflammation of the airways. Asthma affects an estimated 334 million people worldwide and its prevalence continues to rise . This is accompanied by a high economic burden, with recent estimates suggesting an annual cost (both from medical costs and loss of productivity) of €72.2 billion in the European Union  and $56 billion in the United States . Asthma is a heterogeneous disorder with numerous immunological pathways contributing to pathogenesis. Historically, allergic asthma has been described as a T helper type 2 (TH2) cell-mediated inflammatory response, characterized by airway and blood eosinophilia. Further investigations have identified a subset of patients with a predominantly neutrophilic asthma characterized by involvement of T helper type 17 (TH17) cells. While asthma can be controlled with corticosteroid treatment, some patients are steroid-resistant and require specific targeted therapies to manage their symptoms. This is more common in those with TH17-predominant asthma .
The use of transcriptomics has been a key method in uncovering the phenotypes of asthma. Microarray analysis of airway epithelial brushings revealed two distinct clusters of TH2-high and TH2-low asthmatic patients, the latter of which displayed no significant difference in expression of TH2 cytokines (IL-4, IL-5, IL-13) when compared to healthy controls . Although there is a systemic component to the pathophysiology of asthma, differentially expressed genes (DEGs) found in the blood of asthmatic individuals may not translate to those found in lung. Results from a recent U-BIOPRED study showed that DEGs associated with eosinophilic TH2 asthma in the lung were not found in whole blood; in fact, only one of the transcripts was detectable (SERPINB2) and not significantly different between asthmatics and healthy controls . Similarly, investigations of PBMCs from asthmatic individuals also displayed a unique gene expression profile differing from that found in asthmatic lung as well as whole blood [7, 8]. Thus, in order to determine blood biomarkers of asthma, it is evident that the merits and limitations of investigating either blood compartment require thorough evaluation. In this study, we sought to compare the gene expression profiles of whole blood and PBMC compartments in a cohort of mild allergic asthmatic individuals and healthy controls.
Characteristics of subjects analysed for both whole blood and PBMC gene expression
26.6 ± 8.2
33.8 ± 15.6
FEV1 (% predicted)
102.1 ± 13.0
90.2 ± 14.2
Methacholine PC20 (mg/ml)
Top 6 upregulated differentially expressed genes between PBMC and whole blood ordered by log2 fold change (log2FC)
Pearson correlation of RNA expression values between whole blood and PBMCs within subjects
Differentially expressed proteins between allergic asthmatics (n = 8) and healthy controls (n = 7)
Unraveling the complexity of human disease has been enhanced with the rapid development of high-throughput molecular profiling. The examination of immune cells within blood samples is a commonly used method to elucidate disease mechanisms . Here, we compared two types of blood sampling commonly used in immunological studies: PBMCs and whole blood.
Our results show that RNA extracted from whole blood stored in PAXgene tubes yielded higher counts of the 730 measured transcripts compared to RNA from PBMC lysates. Only 6 transcripts were expressed in greater quantities in PBMC lysates compared to whole blood, whereas whole blood had 704 transcripts with higher expression. A reason for this may be the lack of neutrophils, basophils, and eosinophils in PBMC samples. Basophils and eosinophils are only a small subset of all immune cells (0–2% and 1–7%, respectively). Neutrophils make up a majority of circulating nucleated blood cells (45–75%), but a relatively low RNA content relative to normal cells . While this may be partially mitigated through overrepresentation of genes important in immune functions within our 730 measured transcripts (almost half of the 64 uniquely expressed transcripts we identified in whole blood are expressed by neutrophils), the strong correlation observed between PBMC and whole blood samples from the same subject (mean 0.87) suggests that granulocytic involvement is unlikely to explain the higher signal obtained from whole blood. We also cannot rule out technical differences in sample processing. Namely, there may have been disparities in the quality of total RNA extracted from the PAXgene RNA preservation reagent compared to the RLT buffer used for PBMC lysates. Within sample types, whole blood samples (r = 0.97) appear to be more robust than PBMC samples (r = 0.90), though both are acceptable as high positive correlations .
The data generated from our study present a contrast to previous studies comparing gene expression profiles obtained from PBMCs and whole blood. Feezor et al., Debey et al., Palmer et al., Bondar et al., Min et al. [13, 14, 15, 16, 17] found that RNA extracted from PBMCs had a higher abundance of gene expression compared to whole blood (PAXgene) as measured through microarray technology. This is likely due to the abundance of globin transcripts present in whole blood (80–90%), which causes decreased sensitivity in detection of other transcripts . Hemoglobin was identified in a majority of these studies as being highly expressed in the whole blood gene expression profiles, and four of the five studies acknowledged globin as a potential cause of low signal-to-noise ratios in whole blood. Our targeted measurement approach to gene expression (730 genes), though more limited in scope, mitigates these effects as globin transcripts were not measured on the NanoString panCancer Immune Profiling Panel. Despite this, correlations between expression profiles of PBMC and whole blood were found to be between 0.78 and 0.91, which are in line with our results [13, 16].
Interestingly, another group used a similar approach to the aforementioned studies, yet found that whole blood (PAXgene) transcripts produced a greater signal compared to PBMC transcripts despite globin (HBB) being highly expressed . A reason for this may be the use of the Affymetrix Human Exon (HuEx) 1.0 ST microarray, which measures gene expression through exon probing and is unique when compared to the aforementioned studies which used the U133A (Affymetrix, Santa Clara, CA), Lymphochip, HumanHT-12 v4 (Illumina Inc, San Diego, CA), and Sentrix Human-6 v2 (Illumina) microarrays. While the HuEx technology is able to discern alternatively spliced transcripts and thus gene isoforms, it has worse detection rates compared to conventional microarrays and has poorer reproducibility for genes with fewer exons . With this in mind, it is possible that the effects of globin were dampened as the globin genes HBA1, HBA2, HBB, and HBD each contain 3 exons, which is low when compared to the mean number of exons per gene (8.8) .
The upregulated transcripts in PBMC samples (EGR1, IL32, FOS, CCL3L1, IFNL1, and EGR2) have distinct roles in lymphocytes and monocytes. FOS, which encodes for the proto-oncogene c-Fos, acts as a suppressor of the immune system . IL-32 is primarily expressed in lymphocytes, monocytes, and natural killer cells, and induces production of other pro-inflammatory cytokines such as IL-8 and TNF-α . IFNL1 encodes for interferon lambda (IFN-λ, also known as IL-29), which is produced in response to viral infections by myeloid cells . CCL3L1 is produced by PBMCs as a monocyte chemoattractant .
The early growth response transcription factors EGR1 and EGR2 play opposing roles in immunity; while EGR1 positively regulates T and B lymphocyte activation upon antigenic stimulation, EGR2 is involved in T cell anergy and is involved in the FasL-mediated apoptotic pathway . In our analysis, we found that both EGR1 and EGR2 were only detected in PBMC samples, and EGR2 was differentially expressed in our asthmatic cohort. EGR2 has been shown to be a crucial mediator of allergic asthma, as its expression is necessary in mast cells to direct CD4+ T cell migration to inflamed lung . Though EGR2 negatively regulates T-cell activation, it is required for TH1, TH2, TH9, TH17, and cytotoxic T lymphocyte (CTL) differentiation and cytokine production . TH2-mediated immunity was also represented in the uniquely expressed PBMC transcripts, as the GATA3 transcription factor and IL-4 cytokine were detected in PBMC but not whole blood samples, and both are crucial in the differentiation of naïve helper T cells towards a TH2 orientation.
Many of the statistically significant transcripts in the asthma whole blood cohort were related to cytotoxicity and natural killer (NK) cells. Differentiation of cytotoxic T-lymphocytes (CTLs) requires TBX21, RUNX3, EOMES, and STAT4 , all of which are associated with TH1-mediated immunity and were significantly downregulated in addition to the NK surface markers KIR2DS1, KIR3DS1, KLRC1, KLRD1, KLRG1, KLRK1. Transcripts encoding for cytolytic proteins such as PRF1, GZMB, and GZMM were all found to be significantly downregulated in the asthmatic cohort. Although CTLs and NK cells are both mononuclear cells, the differential expression of these transcripts was not detected in PBMCs. CD9 was also upregulated, and it has been used as a surface marker of eosinophils in asthma . Overall, the differentially expressed transcripts found in whole blood describe a downregulation of TH1 differentiation and cell-mediated immunity. For a list of gene/protein abbreviations and their official HUGO names, please see Additional file 4.
The squared correlation coefficient of RNA transcript expression to protein expression has been reported to be ~ 0.4, meaning that mRNA levels account for roughly 40% of the variation in protein concentrations . In our 30 measured proteins extracted from PBMCs, we calculated a squared coefficient of correlation of 0.12 between gene transcripts and proteins. However, we identified 12 proteins with strong correlations to their RNA abundance, which is consistent with a gene-specific level of protein expression .
The analysis of differential gene and protein expression identified an interesting insight into the pathophysiology of asthma. The two differentially expressed proteins in our asthmatic cohort were CD28 and OX40 (TNFRSF4), both of which are critical co-stimulatory molecules during T-cell activation. CD28 knockout mice do not develop an inflammatory response during allergen challenge , and OX40 is essential for optimizing CD4 and CD8 T-cell responses upon activation . In our gene expression data, the ligand for OX40 (OX40L, TNFSF4) was also significantly upregulated in the whole blood samples of our asthmatic cohort. The OX40/OX40L axis is a central player in promoting TH2 polarization of naïve T-cells in the lymph node, with knockout mouse models of either molecule showing a marked reduction in airway hyperresponsiveness, eosinophilia, and pulmonary inflammation when challenged with ovalbumin [34, 35]. OX40 is expressed on T-cells, but can also be found on NK cells, NKT cells, and neutrophils , whereas OX40L is primarily expressed on antigen presenting cells such as B cells, macrophages, and dendritic cells but can be induced on other cell types such as mast cells , basophils , NK cells , and neutrophils . Given the predominant role of TH2 cells in the mechanisms of asthma, it is no surprise that its activator OX40 is found to be upregulated in PBMCs of asthmatics and that widespread OX40L gene expression from a variety of cell types can be detected in whole blood, suggesting a consistently upregulated phenotype of T-cell activation.
In summary, we show here that PAXgene-preserved whole blood is excellent for producing gene expression data with minimal variability and good sensitivity compared to PBMC samples. Despite the small sample size of our study, our paired analysis of whole blood and PBMC samples increased our power to detect differences. Studies requiring blood RNA samples across multiple sites should consider RNA from whole blood instead of PBMC due to its ease of processing (obtained directly from blood draw vs Ficoll isolation) and storage. Although we were limited by the number of transcripts examined, our gene panel did not contain globin genes and thus our data was not hindered by the suppressive effects of globin on the measurement of other transcripts. An added benefit of utilizing the NanoString platform is its precision and robustness; previous studies examining its reproducibility have found that it is comparable to quantitative PCR (qPCR) and exceeds that of RNA-seq and microarrays [41, 42, 43]. For the first time, we have shown gene expression profiling differences between blood compartments when comparing asthmatic subjects with non-asthmatic healthy controls. Furthermore, in whole blood transcript and PBMC protein expression, we identified the upregulation of the OX40/OX40L axis in the peripheral blood of asthmatic individuals, which has previously been identified as upregulated in serum of pediatric asthmatic patients [44, 45] and investigated as a potential therapeutic target in asthma .
The authors thank the research participants for volunteering to take part in these studies, as well as Christopher DeHeer (NanoString) for technical assistance.
J-PO and GMG recruited and collected participant samples; CXY, BS, and SJT designed and performed experiments; DH and CXY analyzed the data; DH, AS, CPS, J-PO, GMG, and SJT discussed results; and DH and SJT wrote the manuscript. All authors read and approved the final manuscript.
Support for this project was provided by AllerGen NCE and the BC Lung Association, as well as an unrestricted grant-in-aid from NanoString for the Vantage 3D assays.
Ethics approval and consent to participate
All study subjects provided written, informed consent prior to sample collection, and the study was approved by the research ethics boards of McMaster University and the University of British Columbia.
Consent for publication
The authors declare that they have no competing interests.
- 2.Gibson GJ, Loddenkemper R, Sibille Y, Lundback B. The European lung white book: respiratory health and disease in Europe. Sheffield: European Respiratory Society; 2013.Google Scholar
- 12.Hinkle DE, Wiersma W, Jurs SG. Applied statistics for behavioural sciences. Boston: Houghton Mifflin; 2003.Google Scholar
- 13.Bondar G, Cadeiras M, Wisniewski N, Maque J, Chittoor J, Chang E, et al. Comparison of whole blood and peripheral blood mononuclear cell gene expression for evaluation of the perioperative inflammatory response in patients with advanced heart failure. PLoS ONE. 2014;9(12):e115097.PubMedPubMedCentralCrossRefGoogle Scholar
- 14.Debey S, Schoenbeck U, Hellmich M, Gathof BS, Pillai R, Zander T, et al. Comparison of different isolation techniques prior gene expression profiling of blood derived cells: impact on physiological responses, on overall expression and the role of different cell types. Pharmacogenom J. 2004;4(3):193–207.CrossRefGoogle Scholar
- 20.Sakharkar MK, Chow VT, Kangueane P. Distributions of exons and introns in the human genome. Silico Biol. 2004;4(4):387–93.Google Scholar
- 37.Kashiwakura J, Yokoi H, Saito H, Okayama Y. T cell proliferation by direct cross-talk between OX40 ligand on human mast cells and OX40 on human T cells: comparison of gene expression profiles between human tonsillar and lung-cultured mast cells. J Immunol. 2004;173(8):5247–57.PubMedCrossRefPubMedCentralGoogle Scholar
- 41.Veldman-Jones MH, Brant R, Rooney C, Geh C, Emery H, Harbron CG, et al. Evaluating robustness and sensitivity of the NanoString Technologies nCounter platform to enable multiplexed gene expression analysis of clinical samples. Cancer Res. 2015;75(13):2587–93.PubMedCrossRefPubMedCentralGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.