It could be low quality/quantity or a problem with the efficiency of the RT-PCR. There could be degradation of your targets before the RT-PCR is performed. Check your protocol to make sure you take all the precautions to avoid RNA degradation. To get consistent results in every run with RNA is the most difficult part. Good luck!
Hi Ratih, it can depend on the quality of each individual RNA and on its quantification. There are two ways of evaluating quality of your RNA.
The first is chemical quality, and this you can do using a UV spectrophotometer, better if you have a scanning one that goes from 200 to 320 such as a NanoDrop or NanoVue. You will have to look at two important absorbance ratios: 260/280 nm and 260/230 nm. The 260/280 nm ratio for RNA at neutral pH (therefore in 10 mM Tris/HCl pH 7.4) should be between 1.9 and 2.1 (in water RNA will be acidic and the ratio will be lower, making it more complicated to spot possible contamination). If you have a lower ratio, it means you have contamination from phenylated compounds, these can be proteins (which contain phenylalanine, tyrosine, triptophane) or phenol (if you used it for the extraction), since they absorb at 280 nm. This will indeed produce variability in your further processing of the RNA, mostly in the efficiency of your reverse transcription reaction, since easily among the protein contamination you will have some RNAse activity affecting the RNA during the RT reaction and if you have phenol this is a denaturing agent that will variably inactivate the RT. The second ratio, 260/230 nm, should always be >2.2. If it is lower it means you have double bond contamination, typically aldehydic and ketonic groups, such as you find in sugars (carbohydrates), or imines, carbon-nitrogen double bonds, such as in guanidine hydrochloride or thiocyanate, present in Trizol reagent and similar. If you have not used Trizol or similar reagent to extract your RNA, and you have a low 260/230 nm absorbance ratio, you probably only have some sugars or pigments in the RNA and this should not interfere with your further processing of the RNA. Finally, the UV absorbance reading at 260 nm will give you a concentration, but ALL nucleic acids as well as dNTPs and NTPs will contribute to that absorbance, so the UV reading of the extracted RNA will not be a realistic quantification of your intact extracted mRNA and is a highly contributing factor to your final variability.
The other part of evaluating the quality (and quantity) of the RNA prepared is running an agarose gel (even a standard TAE 1.5% gel is enough) and evaluating by nucleic acid fluoescent staining (with ethidium bromide or similar dyes) the intensity ratio between the 28S and 18S ribosomal RNAs: since their size ratio is about 2.5:1, you should have a ratio of at least 2 on your gel. An even better approach, if you have the possibility, is to run your RNA on a BioAnalyzer (Agilent) which quantifies the RNA after chromatografic separation by 260 UV absorbance. If the ratio between 28S and 18S is different from sample to sample, it means the degradation degree is different in each sample and this will be reflected in your final RT-PCR product.
For this reasons, in RT-(q)PCR experiments some genes considered to be biologically invariant in the different samples (calibration genes, such as GAPDH, TBP, rRNA, ribosomal proteins, etc) are used to control for extraction and RNA quality problems, calibrating the quantities obtained for your genes of interest (target genes).