Establishment of a trace metal monitoring program and process control parameters are critical for the consistent manufacturing of biotherapeutics. An important part of this monitoring program is establishment of a robust analytical method to detect and quantify trace metal impurities. The guidance in the US Pharmacopeia chapter USP<1023>, “Evaluation strategy for trace elements in cell culture media used in the manufacture of recombinant therapeutic proteins,” provides direction for manufacturers and suppliers seeking to smoothly navigate this aspect of regulatory compliance. To that end, and to support biomanufacturers and process developers in designing robust and consistent processes, we've developed in-house methods to test raw materials and complex cell culture media that fully align with guidelines listed in USP<1023>. These test methods accurately and precisely detect metals in raw materials and complex cell culture media products and represent a cornerstone of our raw material characterization program.
Trace metal monitoring in raw materials and cell culture media
Trace metals in complex cell culture media can have a significant impact on critical quality attributes of recombinant therapeutic proteins. It has been well documented that trace metal concentrations directly impact cell growth, protein titer, and protein quality attributes such as protein glycosylation, charge variants, or protein aggregation (1).
Multiple sources contribute to the trace metal content in cell culture media, including product packaging, manufacturing processes, and raw material impurities (2,3). Raw materials are perhaps the greatest sources of trace metals and are therefore the primary focus of efforts to control trace metals in cell culture media. Many complex cell culture media products have more than 70 components, each bringing potential impurities and adding to the overall metal profile. This additive effect necessitates accurate quantitation of low-level impurities across a wide variety of raw materials and cell culture media products.
Methods for quantitation of trace metals
Historically, inductively coupled plasma mass spectrometry (ICP-MS) has been favored for trace metal analysis due to its high sensitivity, detection, and quantification capabilities. But developing methods for ICP-MS to detect and quantify trace metal impurities in raw materials and complex cell culture media products can be challenging. One significant challenge is that of matrix effects, which occur when concentrations of one or more analytes bias or affect measurements of other analytes. For example, the presence of easily ionizable elements such as sodium, potassium, calcium, and magnesium, which are commonly used in high concentrations in cell culture media, can drastically affect testing results for iron, copper, and manganese (4).
The methods we describe here account for such challenges and fully align with guidelines and USP<730>. We use instrument controls such as continuing calibration verification (CCV), matrix spike recovery samples, and internal standard samples to identify and correct for testing bias due to instrument drift and matrix effects. We also use techniques such as matrix matching, material dilution, material digestion, and internal standard compensation to compensate for measurement bias.
Risk scoring of raw materials for trace metals
An important step in monitoring and controlling trace metals in cell culture media is to develop a raw material risk score system (5). At Cytiva, we use several factors to categorize and rank raw materials by risk, including risk of trace metal impurities. For this, we assess criteria listed in USP<1023>, including material source, material compendial grade, manufacturing process, material complexity, quantity of raw material used, and supply chain considerations (Table 1).
Table 1. USP<1023> Trace metal risk scoring criteria applied to a raw material. Criteria are all weighted evenly in this example, but weighting certain criteria is a common approach.
| Risk ranking criteria | Numeric score | Maximum total risk | Notes |
| Material source | Mined Material - 1.0 (High Risk) Animal/Plant Sourced - 0.66 (Medium Risk) Synthetic Source - 0.33 (Low Risk) |
1.0 | |
| Compendial grade | Noncompendial Material - 1.0 (High Risk) Manufactured to Single Compendial Monograph - 0.66 (Medium Risk) Manufactured to multiple compendial Monographs - 0.33 (Low Risk) |
1.0 | |
| Manufacturing process | Mineral Derived - 0.75 +MCS (High Risk) Fermentation/Purification = 0.5 +MCS (Medium Risk) Chemical Synthesis = 0.25+MCS (Low Risk) |
1.0 | MCS = Manufacture Consistency Score |
| Material complexity | Multicomponent/Chemically Undefined -1.0 (High Risk) Single Component/Chemical Undefined - 0.80 Single Component Large Molecule - 0.60 (Medium Risk) Complex Small Molecule - 0.40 Simple Small Molecule - 0.20 (Low Risk) |
1.0 | |
| Raw material qty | Weight % of Component Used in Product Formulation | 1.0 | |
| Supply/change traceability | Percent Available Change Control | 1.0 | Defined by assessment of: 1. Disclosure of Manufacturing Location 2. Disclosure of Manufacturing Process 3. Disclosure of Material Characteristics and Properties 4. Availability and Suitability of Material Packaging, Labeling, and Documentation 5. Disclosure of Manufacturing Equipment 6. Disclosure of Material Test Method and Acceptance Criteria 7. Level of Supply Chain Disclosure 8. Assessment of Change Control Support |
| Overall risk assessment score | Calculated as the Sum of Material Source, Compendial Grade, Manufacturing Process, Material Complexity, Raw Material Qty, and Supply Change Traceability | 6.0 |
Scores in each category are combined to generate an overall risk score for each raw material, enabling comparison across materials. Figure 1 shows an example comparison of ten different raw materials. Note that no single attribute consistently correlates with the overall risk score: the composite score for each raw material is composed of a unique combination of risk attributes. This finding suggests that reliance on only one or two attributes would be insufficient to accurately capture impurity risk for a raw material.
Fig 1. Comparison of composite trace metal risk scores for ten different raw materials (RM1–RM10). Scores above 3 would be subject to more stringent testing requirements.
The final goal of creating a risk scoring system is that the score determines the degree or intensity at which raw materials are monitored. For the examples in Figure 1, all raw materials with an overall risk score ≥3 would have a higher frequency of trace metal testing with potential implementation of more stringent specification requirements.
Trace metal analysis of raw materials
Development of methods to detect trace metals in raw materials can be challenging. The in-house methods that we use comply with the guidelines and standards described in USP<1023> as well as USP<730> (“Plasma spectrochemistry”). Table 2 demonstrates our method capabilities for trace metal detection in a raw material, using ferric ammonium citrate (FAC) as an example. In this example, USP<730> requirements for accuracy and range, precision, limit of quantification, linearity and internal standards are all satisfied. We developed similar methods for additional materials, including those highlighted in USP<1023> (Table 3).
Table 2. Cytiva’s in-house USP<730>-compliant method validation results for ferric ammonium citrate.
| Ferric ammonium citrate USP<1023> and USP<730> Criteria |
||||||
| Accuracy & Range | Precision | Limit of quantification (LOQ) | Linearity | Internal standard | ||
| Active Ingredient | 95.0% - 105.0% Mean Recovery | ±5% RSD | Target LOQ (ppm) | Result | R ≥0.995 | ±20% Recovery |
| Fe | 99.3% | 0.7% | N/A1 | Pass | Pass | |
| Impurities | 70.0% - 150.0% Mean Recovery | ±20% RSD | Target LOQ (ppm) | Result | R ≥0.99 | ±20% Recovery |
| Al | 102.5% | 0.7% | 0.050 | Pass | Pass | Pass |
| Ti | 109.5% | 1.9% | 0.030 | Pass | Pass | Pass |
| V | 103.0% | 0.0% | 0.010 | Pass | Pass | Pass |
| Cr | 104.5% | 3.4% | 0.010 | Pass | Pass | Pass |
| Mn | 104.0% | 2.7% | 0.010 | Pass | Pass | Pass |
| Ni | 99.5% | 0.7% | 0.010 | Pass | Pass | Pass |
| Cu | 100.0% | 1.4% | 0.010 | Pass | Pass | Pass |
| Se | 109.5% | 3.2% | 0.500 | Pass | Pass | Pass |
| Mo | 99.5% | 0.7% | 0.005 | Pass | Pass | Pass |
| Cd | 92.0% | 3.1% | 0.005 | Pass | Pass | Pass |
| Sn | 90.0% | 0.0% | 0.010 | Pass | Pass | Pass |
| Sb | 91.5% | 0.8% | 0.005 | Pass | Pass | Pass |
| Mg | 102.5% | 0.7% | 0.030 | Pass | Pass | Pass |
| Co | 101.0% | 1.4% | 0.005 | Pass | Pass | Pass |
| Zn | 100.5% | 0.7% | 0.020 | Pass | Pass | Pass |
| 1 In alignment with USP<1225> no target LOQ is listed for the active ingredient the material is diluted to ensure that the theoretical iron content will be within calibration range. Reported quantitation limit for this run was 0.07% with an assay value of 17.5%. | ||||||
Fig 2. Comparison of sample preparation effectiveness for trace metal analysis of ferric ammonium citrate. We prepared samples of FAC under four conditions (with or without digestion, with or without matrix correction). The USP required range for matrix spike recovery (70% to 130%) is highlighted in blue. Trace metals are color-coded based on the boxes in the legend on the right.
Table 3. Example raw materials and products with Cytiva’s in-house USP<1023>-compliant methods
| USP<1023> and USP<730> Criteria for method validation | |||||
| Accuracy & Range | Precision | Limit of quantification (LOQ) | Linearity | Internal standard | |
| Specification | Active Substance: 95.0% - 105.0% Impurity Analysis: 70.0% - 130.0% recovery |
Active Substance: ±5% RSD Impurity Analysis: ±20% RSD |
Meets Target LOQ | Active Substance: R ≥0.995 Impurity Analysis: R ≥0.99 |
±20% Recovery |
| Sodium phosphate dibasic | ✔ | ✔ | ✔ | ✔ | ✔ |
| Sodium chloride | ✔ | ✔ | ✔ | ✔ | ✔ |
| Glucose | ✔ | ✔ | ✔ | ✔ | ✔ |
| L-Glutamic acid | ✔ | ✔ | ✔ | ✔ | ✔ |
| L-Proline | ✔ | ✔ | ✔ | ✔ | ✔ |
| Ferric ammonium citrate | ✔ | ✔ | ✔ | ✔ | ✔ |
| Complex cell culture media | ✔ | ✔ | ✔ | ✔ | ✔ |
Trace metal testing of complex raw materials often require additional method optimization. Materials with high organic content, must be thoroughly with high concentration acids in a microwave digestion system employing high temperatures and pressures before analysis to ensure accurate results (Figure 2). Additional correction methods can be employed to mitigate matrix effects. Some of these include matrix matching, material dilution, material digestion, and internal standard compensation. Using these methods, we can secure accuracy and repeatability of ICP-MS data for our raw materials.
Trace metal testing of complex cell culture media
As difficult as developing methods for trace metal detection in raw materials can be, developing methods for complex cell culture media is even more so. We’ve employed several techniques to develop a method that complies with USP<1023> guidelines. As recommended in USP<1023>, this method uses a series of internal standard solutions to monitor and correct for matrix and spectral interference.
Figure 3 shows an example of matrix spike recovery for a complex cell culture media product using our method. As recommended in USP<1023>, we tested 1:6 and 1:100 dilutions of the product. Due to the high levels of organic carbon content in complex cell culture media, we used a microwave digestion procedure to reduce or eliminate matrix effects due to hydrocarbon content. And, because we're investigating elements with concentrations that span several orders of magnitude in complex cell culture media, we used serial dilution to ensure accurate quantitation in the range of the standard curve.
Fig 3. Matrix spike recovery for complex cell culture media using 1:6 and 1:100 dilutions. Matrix spiking was completed using the material calibration standard at a 10 ppb level for the 1:6 dilution (trace impurity analysis) and at 1000 ppb level for the 1:100 dilution (intentionally added component assay method). Method limits were set at ±30% spike recovery for the trace impurity method, and ±5% for the assay method.
Using this method, we were able to detect trace metals in complex cell culture media in the low parts per billion (Table 4).
Table 4. Limits of detection for trace metals in complex cell culture media.
| Element | Al | Ti | V | Cr | Mn | Ni | Cu | Se | Mo | Cd | Sn | Sb | Mg | Fe | Co | Zn |
| Detection limit (ppm) | 0.067 | 0.055 | 0.008 | 0.018 | 0.007 | 0.008 | 0.019 | 0.617 | 0.005 | 0.007 | 0.015 | 0.003 | 0.184 | 0.258 | 0.025 | 0.237 |
Conclusion
At Cytiva, we’ve developed numerous in-house methods for detection and quantification of trace metal impurities in raw materials and complex cell culture media products. These methods align with the USP standards and guidelines described in USP<730> and USP<1023>. Many of these methods employ additional techniques to eliminate bias caused by issues such as matrix effect. Doing so allows for accurate detection of trace metals such as Al, Tl, V, Cr, Mn, Ni, Cu, Se, Mo, Cd, Sn, Sb, Mg, Fe, Co, and Zn at low ppb levels.
RELATED SERVICES
REFERENCES
- BioPhorum. Trace element variation for chemically defined cell culture media: biopharmaceutical industry requirements and cross-company collaboration to mitigate risks. BioPhorum. 2022. https://www.biophorum.com/download/trace-element-variation-for-chemically-defined-cell-culture-media-biopharmaceutical-industry-requirements-and-cross-company-collaboration-to-mitigate-risks/
- Dickens J, Khattak S, Matthews TE, Kolwyck D, Wiltberger K. Biopharmaceutical raw material variation and control. Curr Opin Chem Eng. 2018 Dec;22:236–43.
- Cytiva. Understanding and controlling raw material variation in cell culture media. https://www.cytivalifesciences.com/solutions/bioprocessing/knowledge-center/Cell-culture-material-variability. Accessed 22 August 2024.
- Grindlay G, Mora J, de Loos-Vollebregt M, Vanhaecke F. A systematic study on the influence of carbon on the behavior of hard-to-ionize elements in inductively coupled plasma–mass spectrometry. Spectrochimica Acta B. 2013 Aug;86:42–9.
- Quinn KS, Liu C, Paulakovich Z, Neenan S, Mack A, Moebius P, et al. Raw material risk assessments - a holistic approach to raw material risk assessments through industry collaboration. BioPhorum. https://www.biophorum.com/download/raw-material-risk-assessment-september-2019/. Published 2019. Accessed 20 August 2024.