The limiting factor for obtaining meaningful gene expression is the quality of the initial RNA preparation. RNA purity and integrity are of foremost importance to ensure reliability and reproductibility of qRT-PCR . Although the use of cell culture and laboratory animals allowed quick processing of the RNA under tightly controlled protocols, this is not always the case for human samples. It is especially true, for example, for human placenta obtained immediatly after delivery. In the studies of placental gene expression in uncomplicated pregnancies as well as in pregnancies complicated with diabetes, we were faced with the collection of placentas occuring at any hour of day and night. Thereby, time course studies of RNA expression and degradation seemed to us critically needed in order to evaluate the biostability and quality of placental RNA species, i.e how long a placental tissue may be stored without degradation of RNA. For some studies, the heterogeneity of the placenta tissue requires a previous dissection of tissue to isolate the villosities in order to study specific gene expression . It was of interest to evaluate the effect on RNA yields, integrity and stability on placenta according to handling of the tissue before banking. Therefore, we evaluated the effects of post delivery delay time and tissue handling on RNA integrity and mRNA expression levels.
Reliable statistical analysis of these parameters leaded us to investigate a sufficient panel of tissues. Therefore our study was performed on 140 samples from 14 placentas, 2 protocols of preparation and 5 delay time points. Power calculations for qRT-PCR comparisons usually indicate that a sample size of at least 50 is required to detect difference. Hynd et al, reported that 13 cases by group yielded statistically differences on a range of widely disparate parameters .
The first parameter studied was the pH of placental tissue. pH was of interest in the light of hypoxia related to tissue injury. Hypoxia is associated with an accumulation of lactates, a lower pH and a subsequent activation of acid lysosomial RNAses . There were no consistent differences in tissue pH between placentas whatever the mode of delivery. Tissue pH was found stable at + 4°C up to 72 h. A significant fall of pH was found after 96 h of storage at + 4°C. This stability has been already reported in brain tissue by others [9, 11]. The overall yield of RNA was found in agreement with reports from other studies on placenta . Otherwise, the yields were lower for placentas previously dissected. This suggests an activation of lysosomal RNAses by tissue disruption leading to a degradation of total RNA [7–11].
The absence of significant variation according to delay time shows that degradation of RNA seems to depend more on tissue handling than on delay time of storage. This agrees with several reports performed on various tissues [13–17].
The assessement of total RNA integrity can be done by two main methods: the standard 28S:18S ratio and the recent RIN integration method. The standard method uses electrophoresis of RNA and the evaluation of 28S and 18S bands and ratios. It is commonly accepted that intact RNA has a rRNA band ratio > 1.8 . We reported very stable values in all placentas samples. The more recent method uses capillary electrophoresis and accurate integrations of peaks expressed as RIN (RNA Integration Numbers). RIN ranges from 1 to 10 with 1 being the most degraded profile and 10 the most intact . In solid tissue, (6–8) RIN values are considered as valuable and reliable RNA . Placenta samples dissected before extraction showed lower values than samples quickly treated with RNA later™. This significant decrease of RIN values according to handling accounts for a partial degradation of tissues by dissection. This might be explained by the activation of intracellular RNAses during tissue disruption . This agrees with corresponding RNA concentrations described in Fig. 1. Several studies reported a good correlation between RIN values and qRT-PCR relative units [13, 21]. They recommended then to consider a RIN > 6 for a suitable total RNA and RIN > 8 for a perfect RNA. Strand et al showed that RIN > 6 correlate with suitable expression of various genes while RIN < 6 are associated with a decrease expression of these genes . Therefore, only total RNA recovered from placentas samples extracted without dissection and stored up to 96 h in RNA-later™ may be considered as reliable for qRT-PCR.
Our observations suggest that despite the 28S:18S ratio is considered as the gold standard for the measure of integrity, it lacks precision and discrimination between preserve and partially degraded RNA.
Following total RNA integrity, determination of mRNA integrity is important to assess. We analysed it by the quantification of 5' and 3' ends fragments of some gene transcripts. The fragments located towards the 5'end of the mRNA of housekeeping genes are used as indicator sequences for the degree of degradation [17, 22].
FASN 5'/3' observed ratios are higher than those reported by Bauer in blood samples . This might be explained by a high stability of 5' ends of genes expressed in placenta tissues . FASN and GAPDH 5'/3' ratios are higher in protocol A compared to protocol B and are probably related to a partial degradation of mRNA after dissection. Expression levels of FASN and GAPDH fragments were stable up to 48 h when samples were kept at 4°C. This shows that the delay of storage has an effect on the integrity of mRNA after 48 h. This effect is higher in protocol B and fits with the decrease of total RNA integrity measured by RIN. Previous studies have observed intact total RNA in various tissues stored at + 4°C post mortem, such as human brain (up to 36 h), human bone.(up to 48 h) liver of rabbit (up to 96 h) and bovine muscle (up to 8 days) [15, 17, 24]. This confirms the high stability of RNA in most of tissues when stored at low temperature. The delay time has a less effect on GAPDH expression in samples treated with protocol A. In fact, we obtain very stable ratios compared to those observed with FASN. The difference of lengths of FASN and GAPDH transcripts respectively 8 kb and 3 kb might explain this difference. A very long transcript is more sensitive to partial degradation than a smaller one.
Endogenous controls, usually housekeeping genes, are measured to better normalize between tissue samples [25–28]. The choice of good controls is tissue dependant, and the same housekeeping genes suitable for a tissue are not for another [27–29]. Previous studies have compared a set of housekeeping genes in placenta by qRT-PCR . B2M, ALAS and cyclophilin have been reported as stable genes in placenta and so far used in this study . Our observations highlight variability of expression profiles for these 3 genes according to handling and/or delay time. B2M appears as the most stable gene no sensitive to conditions of storage or tissue handling. This agrees with a previous study showing that B2M is one of the most stable housekeeping gene in placenta . ALAS mRNA expression is sensitive to storage and cyclophilin to tissue handling. This seems not to depend on the length of transcripts that are quite similar for these 3 genes. It is important to note that the length of the amplicon is over 200 bp for cyclophilin and about 100 bp for B2M and ALAS. Others reported a correlation between total RNA integrity measured by RIN and the efficiency of qRT-PCR according to the length of the PCR product [21, 30]. Taken together, these results highlight that storage and handling influence the expression of standard housekeeping genes in placentas. B2M was found the most stable gene in placentas stored up to 48 h whatever the mode of preparation.
The analysis of TNFα and COX2 mRNA PCR products show a stability of expression up to 48 h and thereafter an important up regulation (4 fold and 9 fold at interval 72 h and 96 h respectively). TNFα and COX2 are involved in cellular defense and stress response. Overexpression of these genes induced by storage of rat liver at 37°C has been already reported . The mechanism may be a stabilization of labile mRNA through, for example the activation of MAPK or other signaling pathways . This activation of signaling pathways might be enhanced by ischemia and apoptosis of tissues during a long period of storage. Despite the little effect of delay time of storage on RNA integrity, our observation shows that it is important to take into account these variations of expression of inducible genes.