| | RT-PCR#1 | RT-PCR#2 | RT-PCR#3 | RT-PCR | Microarray |
|---|
Gene | F.C. | P value | F.C. | P value | F.C. | P value | Mean F.C. | Mean F.C.a
| P value* |
|---|
AGXT2L1
| 21.19 | 0.003 | 17.36 | 0.010 | 15.57 | 0.005 | 18.04 | 9.45 | 0.001 |
SPP1
| 5.92 | 0.004 | 6.59 | 0.002 | 8.09 | 0.001 | 6.87 | 4.16 | 0.040 |
ITGAV
| 3.21 | 0.004 | 3.20 | 0.026 | 3.33 | 0.011 | 3.25 | 3.43 | 0.002 |
ZBTB1
| 2.17 | 0.022 | 7.22 | 0.001 | 3.51 | 0.001 | 4.30 | 3.39 | 0.008 |
IFRD1
b
| 1.16 | 0.731 | | | | | 1.16 | 3.39 | 0.001 |
BAMBI
| 2.26 | 0.002 | 1.81 | 0.007 | 2.16 | 0.026 | 2.08 | 3.1 | 0.003 |
IER5
| 4.43 | 0.051 | 4.58 | 0.002 | 4.41 | 0.011 | 4.47 | 2.98 | 0.003 |
ARMET
| 2.64 | 0.002 | 2.94 | 0.001 | 3.04 | 0.001 | 2.87 | 2.87 | 0.004 |
LPIN1
| 2.41 | 0.038 | 2.25 | 0.011 | 2.58 | 0.007 | 2.41 | 2.45 | 0.002 |
DNAJA4
| 5.42 | 0.038 | 4.63 | 0.011 | 4.65 | 0.011 | 4.90 | 2.15 | 0.021 |
ATF4
| 2.31 | 0.004 | 2.56 | 0.002 | 2.07 | 0.011 | 2.32 | 2.04 | 0.003 |
DLC1
| 3.66 | 0.001 | 3.47 | 0.002 | 3.47 | 0.001 | 3.53 | 1.93 | 0.008 |
ANXA5
| −2.39 | 0.002 | −3.15 | 0.001 | −2.99 | 0.001 | −2.84 | −2.52 | 0.033 |
EPHX2
| −3.81 | 0.001 | −2.28 | 0.001 | −2.45 | 0.001 | −2.85 | −2.53 | 0.007 |
KCNJ13
| −2.82 | 0.004 | −2.56 | 0.007 | −2.86 | 0.002 | −2.75 | −2.54 | 0.018 |
CIRBP
| −4.72 | 0.002 | −7.12 | 0.001 | −6.84 | 0.001 | −6.23 | −2.72 | 0.003 |
CLDN5
| −8.80 | 0.001 | −5.49 | 0.001 | −4.99 | 0.001 | −6.43 | −3.1 | 0.004 |
LYPLAL1
| −3.82 | 0.002 | −3.81 | 0.001 | −4.22 | 0.001 | −3.95 | −3.3 | 0.002 |
PRDM16
| −6.50 | 0.002 | −4.17 | 0.001 | −3.90 | 0.001 | −4.86 | −3.62 | 0.004 |
- Individual gene expression (and validation). In order to assess the validity of the RNA expression data obtained by microarray, we performed RT-PCR analysis on cDNA prepared from laser-dissected CPE of the same healthy control (Br0–1) and late stage AD samples (Br5–6) as used for the microarray. We essentially followed the procedure of van Soest SS et al. (2007) [33]: In short, RT-PCR validation was done, in triplicate, on a selection of 17 genes. These genes were chosen on the basis of high expression level (>90%), significantly different expression between late stage AD and healthy control (p < 0.05) and high fold change (FC). Next, we searched for unique and efficient primers in the last 1 kb of the 3′ region the mRNA, as this area was used for the design of the oligo’s on the microarray and the microarray amplification procedure employed the poly a tail. We measured three times the expression levels of the resulting set of 17 genes by RT-PCR and calculated three times the FC and the relevant p values. Table 2 summarizes the findings and compares RT-PCRs with the outcome of the microarray. The significant difference as indicated by the microarray was confirmed for 16 of these genes, whereas the 17th gene did not amplify well in our hands in the RT-PCR, and showed unexpectedly a very low expression, perhaps due to limitations in primer design. Pearson’s correlation coefficient of the FC’s was remarkably high (r = 0.95) and significant (p < 0.01), confirming the gene expression results of the microarray. F C: Fold change in mRNA expression of CPE of AD (Braak 5,6) vs. healthy (Braak 0,1) donors. Pearson’s correlation coefficient between F.C. of micoarray and RT-PCR was significant (r = 0.95, p < 0.01). RT-PCR#1, RT-PCR#2, RT-PCR#3 denote three independent RT-PCRs on 7 human control samples and 7 and AD affected samples. *Some genes are represented with multiple probes on the array. If so, the value shown is the lowest P value. aSome genes are represented with multiple probes on the array. If so, the value shown is the mean F.C. bOur RT-PCR showed for this gene unexpectedly a very low expression, perhaps due to primer design
