To address the concern of whether the overall RNA gene expression can be sufficiently preserved in DBS to answer significant biological questions, we investigated the gene expression of males vs. females acquired from DBS using microarray technology. We identified four statistically significant differentially-expressed genes between groups of male and female subjects. These genes were X inactivation-specific transcript (XIST), lysine-specific demethylase 5D (KDM5D) which is also known as selected cDNA on Y, homolog of mouse (SMCY), uncharacterized LOC729444 (LOC729444), and testis-specific transcript, Y-linked 21 (TTTY21) (Table 1). All genes were sex-related. The expression of XIST was lower in males than in females. KDM5D and LOC729444 were over-expressed in males compared to females while TTTY21 was under-expressed in females compared to males (Figure 1). Although other sex-related transcripts such as testis determining factor (TDF) could be detected, the gene expression was low in general and there was no significant difference between males and females.

Of the 4 genes that we identified, LOC729444 is an unknown gene while TTTY21 is a non-protein coding transcript. On the other hand, XIST and KDM5D are well-documented genes related to sex traits. Thus, we generated scatter plots using gene expression data (normalized probe intensities) of XIST and KDM5D to investigate whether RNAs related to sex is preserved. Our results demonstrated that the expression of XIST were generally higher in females than in males and vice versa for KDM5D. We confirmed that sexes can be distinguished and gender status is preserved in DBS (Figure 2A,B).

We validated the RNA gene expression of XIST, the most significantly differentially-expressed sex-specific gene, using qRT-PCR. In the microarray data, expression signal of XIST was weak in both sexes, but the expression signal in males was significantly lower than in females. The qRT-PCR result reflected this weak signal, in that we found a detectable XIST signal in only 12 of 31 females in our samples (38.7%) (Figure 3A). However, detection of XIST by qRT-PCR was found only once in 75 male samples (1.3%) (Figure 3B) (p < 2.95 × 10⁻⁶; Fisher’s exact test). Our observation confirmed the findings of microarray data that RNA gene expression of XIST is overall lower in males than in females.

It is a fact that high quality RNA is more difficult to acquire and store as it is more easily compromised when compared with DNA in human clinical samples. Therefore, it is logical to ask if any usable RNA can be obtained from DBS which are usually kept at ambient temperature. With the advancement of molecular tools and technologies, we recently established a technical protocol for using unique amplification systems to amplify non-poly A transcripts from compromised RNA and hybridizing them onto microarrays (with 60-mer length oligonucleotide probes) to obtain gene expression data from blood spot samples [15]. However, there remains the concern of whether the overall RNA gene expression can be sufficiently preserved in DBS to answer significant biological questions. To address this concern, we investigated the gene expression of 75 males and 31 females from DBS aged 2.5–15.6 years using microarray technology and evaluated their differentially-expressed genes. We identified 4 differentially-expressed genes which were statistically significant between males and females: XIST, KDM5D, LOC729444, and TTTY21. All four genes are sex-related.

XIST, which is mapped on Xq13.2, is one of those genes involved in X chromosome inactivation [16]. The X-inactivation center (XIC) regulates X-inactivation by counting the number of X chromosomes and contains XIST which localizes exclusively in the inactive X chromosome [17]. Thus, XIST is only expressed in the inactive chromosome X of a female. Our results are in concordance with the above statement, showing significantly lower XIST expression (p < 1.6 × 10⁻¹⁶) in male than in female DBS samples. In addition, its gene expression from qRT-PCR was in concordance with the microarray data. KDM5D, also known as selected cDNA on Y, homolog of mouse (SMCY), is first isolated in the short arm of the mouse Y chromosome [18]. In mouse, its expression can be detected as early as in the 2-cell stage during blastocyst development. In the human, it is mapped on Yq11.222, encoding one human H-Y (histocompatibility Y) epitope called H-Y/HLA-B7, and is proven to be Y-linked in many species [19]. In DBS, we also showed significantly higher KDM5D expression in our male samples as compared to our female samples (p < 0.0036). LOC729444 is an uncharacterized gene with currently unknown functions. It can be found in the cDNA libraries from testis according to the Cancer Genome Anatomy Project (http://cgap.nci.nih.gov/Genes). In our study, we also found LOC729444 highly expressed in males compared to females (p < 0.042). In addition, we found the expression of TTTY21, a non-protein coding RNA mapped on Yp11.2, significantly lower in females than in males (p < 0.042) from our DBS samples. We also examined the expression of these genes in fresh blood (male n = 102, female n = 36) and frozen tissues (male n = 39, female n = 36) that are available from the public domain. We and others found the expression of XIST significantly higher in females compared with males in fresh blood samples [20]. Generally, a higher intensity and fold-change are also observed in the fresh samples: 4 fold change in XIST (p < 4.7 × 10⁻²⁰) and 3 fold change in KDM5D (p < 2.8 × 10⁻²³), while LOC72944 and TTTY21 showed higher intensities but no significant fold-change (p < 0.022 and p < 0.144, respectively). Sex-specific genes including XIST and KDM5D were also reported to be differentially-expressed in fresh frozen male and female human brains [21]. Thus, although the overall intensity/fold-change is higher in fresh samples, gene expression changes of sex-specific transcripts in DBS are in concordance with those from fresh blood and frozen tissue samples in general. High-quality RNA extracted from matched whole blood at the time when DBS was collected could be used as a reference but unfortunately such samples are difficult, if not impossible to obtain, especially for very old DBS samples.

Overall, we showed that the expression of sex-related genes can be detected, measured, quantified, and correlated with their corresponding expression in each gender. The observed RNA gene expression and gender status preservation in DBS is confirmed by qRT-PCR. Preservation of gender status in DBS is both evidence of and an internal control for accurate gene expression related to certain traits that are preserved in DBS. Consequently, gene expression profiling can be performed to obtain genetic signatures to answer biological questions related to traits of interest.

Archived dried blood spots (DBS) were obtained with parental consents from the Michigan Neonatal Biobank (MNB) which stores and manages Michigan’s four million residual newborn dried blood spot cards (http://mnbb.org/). DBS from 53 pairs of cerebral palsy/healthy control subjects (75 male and 31 female) that had been archived from 2.5–15.6 years at ambient temperature were sampled for a large molecular epidemiological study of cerebral palsy using microarray technology. This study was approved by the Institutional Review Boards of Michigan State University, the Michigan Department of Community Health, and the Van Andel Research Institute.

Total RNA was prepared according to our modified protocol [15]. In brief, three 3 mm punches from each blood spot were homogenized before total RNA extraction using the illustra RNAspin Mini RNA Isolation Kit (GE Healthcare, Buckinghamshire, UK). Total RNA was then concentrated with an RNA Clean & Concentrator-5 Kit (Zymo Research, Orange, CA, USA). RNA quality and quantity were evaluated using an RNA Pico Lab Chip on the Agilent BioAnalyzer (Agilent Technologies, Santa Clara, CA, USA). The average RIN (RNA Integrity Number) was 2.3 ± 0.71. The WT-Ovation Pico RNA Amplification System (NuGEN Technologies, San Carlos, CA, USA) based on the Ribo-SPIA technology was used to amplify total RNA and synthesize first and second-stranded cDNA [22]. Unlike most standard amplification approaches which are based on T7 polymerase cRNA synthesis, this amplification system is initiated both at the 3′ end and randomly throughout the whole transcriptome. Hence, it has the advantage of amplifying non-poly A transcripts and compromised RNA from samples such as archived blood spots in our study.

cDNA was labeled with Alexa Fluor 3 fluorescent dye from the BioPrime Total Genomic Labeling System (Invitrogen Life Technologies, Carlsbad, CA, USA) before purification using the PureLink PCR Purification System (Invitrogen). Purified labeled product was then applied onto a 8 × 60 K whole human genome gene expression microarray (Agilent) for hybridization using standard Agilent procedures. After hybridization for 17 h at 65 °C and 10 rpm rotation speed, the arrays were washed for 2 min each with wash buffer 1 and 2 and scanned with an Agilent G3 high-resolution scanner. Probe features were extracted from the microarray scan data using Feature Extraction software v. 10.7.3.1 (Agilent).

All procedures for data processing and analysis were done in R version 2.13.2. Poor-quality signals on the Agilent microarrays were removed when the gProcessed signal was less than two times the gProcessed signal error [23]. Data were normalized between arrays using quantile normalization method [24–26]. Normalized data were then aggregated to gene level to assess differential expression of all individual genes between males and females by moderate t-statistic from linear model and empirical Bayes method implemented in the limma R package [25,26]. Significance of genes were adjusted for multiple testing using the Benjamini and Hochberg method [27]. Statistical significance was defined when adjusted two-sided p ≤ 0.05. The median average of the expression values from the females g(f) and males g(m) for each gene were computed to generate the heatmap. The group-wise average, [g(f) + g(m)]/2, was subtracted from each expression value to generate relative expression values for plotting.

To validate the microarray results, the most significantly differentially-expressed sex-specific gene, XIST, was evaluated using qRT-PCR using Hs01079824_m1 primer/probe set (Applied Biosystems, Carlsbad, CA, USA). In brief, 20 ng cDNA was mixed with Taqman PCR Super Mix (Applied Biosystems, Carlsbad, CA, USA) and gene-specific forward/reverse primers and probes before subjected to real-time PCR quantification using the ABI PRISM 7500 Fast Real-time PCR System (Applied Biosystems, Carlsbad, CA, USA). The data are presented as number of cycles to threshold.