REGULATORY REQUIREMENTS FOR SYNTHETIC PEPTIDES
Information in the European Pharmacopeia (EP) (1), as summarised in Table 1 highlights the required impurity control for synthetic peptides as they proceed through clinical development.
In general terms, control of impurities should increase through clinical development. Specifically, individual impurities should not exceed their values as qualified in toxicological studies, or be greater than 1.0%, whichever is higher. However, these limits apply to a specific analytical method and as method development is performed, this can result in changes to the levels of impurities previously thought to be qualified. Hence, it is important to perform analytical development with this in mind, and ensure knowledge of the overall impurity profile of a product is understood right from the start.
Analytical development and impurity detection
In analytical development, a screening approach is applied using material generated early in process development. Whilst no two peptide methods are alike, they typically encounter the same challenges:
- Resolution
- Purity of the main peak caused by co-eluting impurities
- Methods with long run times and limited solution stability
Screening, performed on an LC-MS instrument, takes ~3 days and results in the basis of the analytical method that will be further developed and validated. The mass spectrometry information allows the process and analytical teams to understand early if co-eluting impurities are present, which will be resolved analytically at a later date, and thus need control during the process development phase.
Purification Process Development part 1 – screening for selectivity
In a similar manner, purification process development also starts with a screening phase and will typically draw on the information gained in the analytical labs. Whilst the preparative chromatography used in manufacture will have a particle size of ~10µm due to the back pressure limitations of the equipment, screening is typically performed on smaller particle size columns. The main focus on this phase of work is to identify the conditions that offer the best selectivity, since this offers the single biggest opportunity for resolution, and thus impurity control, as shown in the equation below (2).
Looking at each parameter in turn, it’s clear that selectivity (α in the equation above) is the predominant factor in terms of impurity resolution. Furthermore, increasing k (the capacity factor) requires an increase in solvent consumption and an increase in N (column efficiency) is impractical as this relies on a decrease in particle size out-with the limitations of the bulk media used for preparative chromatography.
In order to screen for selectivity even for preparative chromatography, small U(H)PLC columns are used in an analytical setting, with up to 100 chromatograms generated within 24 hours. The media used in this screening phase is typically 1.7µm, but importantly is also available in bulk 10µm for manufacture. The typical parameters that are screened are media, buffer, pH and organic modifier. The chromatograms generated are assessed based on the following criteria, with an appropriate weighting applied to each:
- Purity (lower is better)
- Peak width (lower is better)
- Retention factor of peak before main peak (higher is better)
- Retention factor of peak after main peak (higher is better)
The data is then sorted to provide a graphical output and allow selection of the optimum media and buffer for the next phase of development.
Purification Process Development part 2 – optimisation
After selection of the best conditions for selectivity, attention switches to the optimisation of the purification method, now using 10µm bulk media. The output from a preparative HPLC step is usually a balance between purity, loading and recovery, particularly as most of the impurities generated in the manufacture of peptides have similar chromatographic properties. As can be seen in Figure 4, when incomplete impurity resolution is obtained, which is common in peptide purification, one can either pool all of the product containing fractions to increase the yield and lower the purity, or pool the fractions free from the impurity to get a higher purity but lower yield. Since purity is typically fixed for clinical and commercial peptides, yield is usually compromised. Newer technologies like multi-column counter-current solvent gradient purification (MCSGP) allow for the automated recycling of the mixed fractions which can increase the yield (3).
Aside from having the optimum selectivity in purification, loading is another important factor that has a significant impact on purification. Considering Figure 4, increasing the loading can simply lead to more of an overlap of the impurity and product and hence no overall improvement. However, working with a low loading that improves separation can lead to an impractical chromatographic step in manufacture, potentially requiring hundreds of injections and weeks of manufacture time to purify a batch of crude peptide.
Another key aspect in the optimisation of a purification method is the impact of pH and the potential for orthogonality that this brings. Preparative chromatography relies on significant overloading of the product onto the silica, resulting in isothermal behaviour which is either Langmurian or anti-Langmurian (4). The difference in this isothermal behaviour can be influenced by pH and can be exploited to aid with impurity removal. Figure 5 shows a plot of a peptide versus a key impurity during a purification of a peptide on C18 silica using 0.1% TFA as the modifier. At this pH (pH 2.5), a Langmurian isotherm is in operation leading to the tailing of the product peak. In this case, the impurity essentially elutes throughout the product elution making removal of it impossible.
By switching the pH of the purification to ~ pH 5, the isothermal behaviour of the product changes to anti-Langmuarian, leading to a fronting peak. Since the impurity behaviour does not change significantly, removal of it is now possible (Figure 6).
The final piece of optimisation typically involves the gradient. As in the loading, one can apply a very long, slow gradient to try and separate an impurity, but this can lead to an impractically long preparative purification. Typically, a purification run time should be one hour or less.
Purification Process Development part 3 – final checks
The purpose of preparative chromatography is to deliver a peptide of appropriate quality suitable for either clinical or commercial supply. Even if a purification delivers the peptide of the desired quality in solution, things can still go wrong! In particular, hydrolysis of peptides containing a glutamine or asparagine in the sequence can undergo deamidation to form the corresponding glutamic acid or aspartic acid impurities, or peptides containing an Asp-Pro in the sequence can also undergo hydrolysis, which in either case can lead to a peptide failing specification if the rate of formation is not understood and controlled.
Another aspect that must be controlled is the potential for a peptide to precipitate and/or aggregate in solution. One of the final steps in a peptide purification process is filtration of the final product solution through a 0.22µm filter to remove any bioburden present ahead of lyophilisation and packaging. If a peptide does precipitate from solution, this negates the possibility of performing this critical step and part of the process development phase is to ensure this does not happen under the parameters of the purification process.
Purification scale-up
Process development of peptide purifications is performed in such a way that scale-up ought to be straightforward! The columns and media used in development mimic that to be used in manufacture with a fixed bed height , meaning the scale-up typically involves increasing the loading and liner flow-rate in line with the square of the radius of the column to be used. An example of this is shown below:
As discussed in the previous section, the purification of large quantities of a crude peptide can run for up to 2 weeks and therefore it is critical that the stability data is understood to avoid degradation of the now purified peptide!
Purification process validation
As products move through clinical development and ahead of commercial supply, the process development team will challenge the purification step as part of readiness for process validation. This typically involves statistically designed experiments to determine both the normal operating ranges (NOR) and the proven acceptable ranges (PAR). The use of DoE allows for the interpolation of parameters that interact with each other and combine to create an effect on the purification performance, with an example shown in Figure 7.
In this case, an impurity (RRT 0.967) has been demonstrated to be under full control within the parameters studied (TFA content of the crude peptide and TFA content of the aqueous buffer to be used in purification) with levels of the impurity only varying between 0.146% and 0.152% against a specification limit of 0.35%.
As peptide purifications typically employ a number of repeat injections, this is also studied in development ahead of process validation to understand how many injections would be possible with a single charge of purification media, specifically looking at a change in column performance or significant impurity or product carryover between injections.
Lack of impurity control in purification
In some cases, despite best efforts in purification, an impurity may be impossible to control in purification and must be controlled at source. Hence, identification of which impurities this applies to in a peptide is critical information. An example where this has been performed is shown in Figure 8 below.
Classifying the source of impurities allows for the process development team to focus on the impurities that are difficult to control during purification and hence must be controlled upstream. Impurities identified as degradants in the example above are those that form during the downstream process (e.g. deamidation), and control through careful pH and temperature in the process is typically employed.
Bespoke solutions are usually required to control specific impurities formed during synthesis. Such solutions may be through the judicious choice of coupling/ deprotection conditions/ inclusion of a dipeptide or choice of scavengers used in the cleavage/ deprotection step. In other examples, control of an isomeric impurity is possible by changing the conditions used for coupling of a specific amino acid, as was the case in this example. In other examples, impurities that have formed under cleavage conditions have been controlled simply by the addition of an alternative scavenger.
CONCLUSIONS
The use of peptides as therapeutics is now firmly established in the pharmaceutical industry, with the majority being manufactured using fully synthetic processes. The chemistry used to make peptides is well understood and has resulted in regulatory bodies viewing peptides as quite different from biologics with an increased expectation of control of the process and impurity profiles. This, coupled with increased scrutiny of peptides through advances in analytical technologies, has led to the requirement for process development teams to fully understand how individual impurities are formed and how these may be controlled in the manufacture at an earlier stage of clinical development. Despite the lack of technological advancement in preparative HPLC, there are a number of options available to control impurities during purification, with an early assessment of those that cannot be controlled in purification critical to allow for their control at source.
Table 1. Limits for impurities in synthetic peptides as listed in EP (Table 2034.2).
Table 2. Purification scale-up factors.
Figure 1. Equation for calculation of theoretical plate (separation efficiency).
Figure 2. Summary of impact of main factors on impurity resolution.
Figure 3. Screening chromatography assessment.
Figure 4. Graphic showing impact of impurity resolution on purification yield.
Figure 5. Product versus Impurity elution at pH 2.5.
Figure 6. Product versus Impurity elution at pH 5.
Figure 7. Typical DoE output.
Figure 8. Crude peptide chromatogram with characterised impurities.
REFERENCES AND NOTES
- EP General Monograph 2034; “Substances for Pharmaceutical Use”, Table 2034.2.
- W.J.Lough and I.W. Wainder Ed. High Performance Liquid Chromatography; Fundamental Principles and Practice, 1996.
- The multicolumn countercurrent solvent gradient purification process; Biopharm International, 2007, 22 (1), pp 42-48.
- Scale-Up and Optimization in Preparative Chromatography: Principles and Biopharmaceutical Applications. Chromatographic Science Series, Volume 88.
