Developed an LC-MS/MS method to quantify small molecules in surrogate matrix, validated by ICH M10

Foreign. Hello everyone, My name is Prajita Pandey and I'm Assistant Director Chemistry at Emory Pharma. Today I'm here to walk you through LC Ms. Ms. Method development to quantitate a small molecule in surrogate matrix and validate per ICS M10 guideline. While conducting LC Ms. Ms. Based bioanalysis studies we we may encounter several challenges, some of which being an interference from endogenous analytes or when a matrix is a rare matrix. Today I'm here to walk you through an approach using surrogate matrix to overcome all of these challenges. What is a surrogate matrix? A surrogate matrix is a substitute matrix used to mimic the characteristics and composition of your biological matrix of interest. There are some key advantages of using surrogate matrix in bioanalysis, one of which is overcoming the matrix challenges and this is by reducing the matrix effect and enhancing method robustness. The other important use of surrogate matrix is minimizing interference from the endogenous components, hence improving the detection limit. It simplifies the validation process when there are rare matrices or your natural matrices are unstable, and it also boosts reproducibility by improving the quality control and ensuring consistent results across batches. In our study today, the biological matrix of interest is vitreous humor and our surrogate matrix is human plasma. Before we move further, let's briefly discuss what ICH M10 bioanalytical method validation is so it provides recommendation for validation of bioanalytical methods and ensures the accuracy, reliability and reproducibility of any analytical method. There are some key parameters to consider while performing the validation per ICS M10 guidelines, some of which listed here are are the ones that we conducted for our study. However, I'm going to walk you through each of these parameters and the importance of these parameters. So now let's look at the overall workflow, starting with your test compound X which is spiked in human plasma and vitreous humor, then moving on to the addition of the internal standard and protein precipitation using methanol. After centrifuging, take your supernatant and run LC Ms. Ms. Analysis on it. The first criteria that we looked into was the specificity and selectivity. Once we conducted our sample preparation, we observed the chromatograms coming out of both blank human plasma and blank vitreous humor. We wanted to ensure that the retention time where test compound X elutes is not present in both of these matrices. As you can see in the left panel you can in both human plasma and vitreous humor. We do not see a Peak for test compound X. In right you can see a chromatogram of how a test compound X would look like when it's present in the sample, confirming the specificity and selectivity of this method. Moving on to a recovery when we perform an extraction in a sample preparation step, it's very important to also evaluate recovery of your method. So per that purpose we spiked our test compound X in both human plasma and vitreous humor and conducted the sample preparation using internal standard and protein precipitation and analyzed it. This was compared to a sample preparation that was performed first. The extraction was performed without the test compound. So after the extraction test compound X was spiked in the sample and LC Ms. Ms. Analysis was conducted on it. The response from the pre spike sample and post spike samples were compared to each other and as you can see on the table, our recovery ranged somewhere from 80 to 98% which are within reasonable limits of bioanalysis. Moving on to the carryover Carryover is another challenge that comes with the Bioanalysis and per ICS M10 guideline, the analyte response at blank should be less than 20% at LLOQ. In this example here you can see that the Response at lloq is 332 and the response at blank is 0, which shows that the analyte response at blank is less than 20% indicating no carryover with this method. Now let's look at the linearity and range. For this purpose we created a calibration curve ranging from 2.5 to 5.2 and this run was repeated over several days using several analysts and as you can see on the table, our accuracy fell within 85 to 115% which is within the ICS MTIN guidelines and also our lower up limit of detection within 80 to 120%. To determine accuracy and precision of this method, we had created quality control samples at four different concentration level at five replicates. For each, we looked at the accuracy coming out of each of these levels and also calculated percent CV to determine the precision. As you can see that our accuracy values fell within 85 to 115% and 80 to 120@LLOQ. Looking at the coefficient of variation of we can see that it's within plus minus 15 and within plus minus 20 at Lloq, which is all within the validation guidelines. Parallelism study so this study is very key when there is a use of surrogate matrix. This determines that is your surrogate matrix actually suitable to quantitate the analyte of interest. So for that purpose we conducted a parallelism study where we compared the response coming out of the quality control samples in human plasma versus quality control samples that were created in vitreous humor. In this example here we have four different quality control samples prepared in vitreous humor. Again, we looked at the accuracy and percent CV. They all were within the acceptance criteria per ICS M10 guidelines. Pooling vitreous humor Another way to measure the parallelism of your study is to actually look at the internal standard response in your human plasma only versus a mixture of human plasma and vitreous humor. In this study here we made a mixture of one part of human plasma and one part of vitreous humor and compared the internal standard response coming out of these samples to the internal standard response only from human plasma. As you can see, the data is very within comparable limits. Also you can see the person's CV is another way of looking into the parallelism is also looking at the internal standard response between your sample and the matrix of interest. Another way of confirming the parallelism of this study of was to look at the instrument response. For that purpose we created a sample that was one part human plasma and one part vitreous humor. And then we prepared other set of sample which was from human plasma only. We compared the internal standard response coming out of these two samples, looked into the average and the person's CV and as you can see in the example here, the internal standard response are comparable. Also the person's CV is within plus minus 15%. One of the other challenges when we are working with bioanalysis is the unknown concentration of the analyte itself. So what do we do when your analyte of interest is in the concentration that's above your calibration curve. For that purpose it's very important to conduct dilution integrity study for that purpose. In this study we we created two stalks at 3,000 nanogram per mil and one 50,000 nanogram per mil in vitreous humor and diluted to make it to target 1500 nanogram per mil and compared the accuracy and percent CV of these samples. As you can see in 2x dilution and 100x dilution. Both accuracy values are within 85 to 115% and the percent CV is within plus minus 50%. Confirming that diluting the sample does not impact the measured concentration of this analyte pre stos stability. This is very important to assess what happens when you remove samples from the frozen storage multiple times. For that purpose we spiked our test compound X in vitreous humor. We put it under three times freeze thaw cycle at minus 80 degrees. We compared the accuracy and made sure that it is within 85 to 115% after three freeze thaw cycles. In this data, as you can see, they were within the acceptance criteria, confirming that even after three freeze thaw cycles the sample were still stable. Bench top stability because there is a sample preparation involved in bioanalysis, we want to make sure that the time it's taken taken to do a sample preparation, the analytes are still stable and we are not compromising with the integrity of the sample. For that purpose, we spiked our sample in both human plasma and vitreous humor kept at room temperature and then we conducted LC Ms. Ms. Analysis to see if the samples were still stable. Because the sample preparation involves several steps, it's very important to ensure that while we are going through those steps, your samples are still stable. For that purpose we took our test compound X spiked in human plasma and vitreous humor. We put it in room temperature for four hours and we conducted LC Ms. Ms. Analysis on it. As you can see, in both human plasma and vitreous humor the accuracy values fell within 85 to 115% of even after sitting in the room temperature four hours, ensuring that it is stable for that duration. While we are preparing our sample before doing the LC Ms. Ms. Analysis. Once the samples are prepared, they are ready to be put in autosampler. But because your sample sequence also has a calibration and QCs, it is very important to ensure that the samples that is sitting in the autosamplers are stable within that time. A sample sequence would have calibration standards, QC samples and the prepared samples in the sequence. While running calibration standard and qcs, your samples are sitting in the autosamplers. Hence it is very important to determine the stability of those samples while sitting in autosamplers to be injected. For that purpose, we created quality control samples at four different concentration level and we calculated accuracy for all of these quality control samples at each of these time points. We conducted this study for 15 hours and we calculated the accuracy values and determined that it were all within 85 to 115%. To assess the long term storage in the freezer, we spiked test compound X in vitreous humor and and stored them at minus 80. We removed the samples as designated time points in this example. Here you can see the data from week one and week three we calculated the accuracy and they all fell within 94 to 115%. This study so far has been conducted up to nine weeks and this study is still going on. As a summary, Today we discussed LCMRM, Ms. method development to quantitate a small molecule in vitreous humor using human plasma as a surrogate matrix. For the purpose of this presentation, we presented you today the vitreous humor data. However, at Emory Pharma we have conducted similar approach to also quantitative the same small molecule in aqueous humor. This approach can be adapted for instances such as rare matrices or if matrices without interference are not available. We also validated this method per ICS M10 bioanalytical guidelines for which we conducted parallelism experiments which demonstrated the suitability of use of surrogate matrix. The surrogate matrix requires greater upfront work, but down the line it streamlines the sample analysis and cost savings. To discuss your project with one of our scientists, please contact BD@mrefarma.com. thank you.