Extending the Visible Residue Limit Approach to Potent Therapeutic Proteins — Part 1: Materials and Methods
Visual inspection for equipment cleanliness is an important element of cleaning validation and monitoring. Multiple regulations, guidances, and standards address expectations for cleaning (1–4). Visual inspection may be used to verify attainment of health-based cleaning limits as described in guidances from the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) (5) and Pharmaceutical Inspection Convention and Pharmaceutical Inspection Co-Operation Scheme (PIC/S) (6) so long as the conditions of Questions 7 and 8 from the associated question-and-answer documents are met (7, 8).
Visual inspection offers several advantages over cleaning verification based on analytical testing (see the “Advantages and Disadvantages” box on the next page) (9, 10). But several published studies have documented widely varying detection limits for visual inspection for small molecules ranging from 0.4 μg/cm2 to >10 μg/cm2 (1, 11–14).
Validation engineers frequently use total organic carbon (TOC) levels as surrogate analytes for assessing equipment-cleaning adequacy, especially for equipment involved in production of recombinant therapeutic proteins. Results from TOC samples should be less than or equal to the acceptance limit (AL) based on the maximum acceptable carryover (MAC) from a previously manufactured product into a subsequent one. Our team performed studies evaluating the viability of using visible residue limits (VRLs) in lieu of TOC values to demonstrate attainment of a carryover acceptance limit. The visible residue limit is defined as the lowest mass (μg) over a given unit area (cm2) that all analysts can see at every angle and lighting intensity.
We performed a small-scale visual-inspection study to determine the VRLs of drug-substance (DS) samples from before and after exposure to a cleaning process. Our goal was to determine whether a VRL approach would be a viable replacement for TOC-swab acceptance limits of <0.7 ppm for native protein and <0.5 ppm for degraded protein. In part 1 of this article, we describe our experimental strategy.
Part 2 (in press for BPI’s June 2024 issue) will present test results and discuss their implications for applications of VRL-based approaches.
Summary of Our Methodology: For postexposure samples, we used degraded protein content to represent the actual state of the protein product (active pharmaceutical ingredient, API) observed on the equipment surface after cleaning. Native protein was used in assessing visible residue before degradation. Our team tested three therapeutic-protein products representing three biological modalities: a monoclonal antibody (mAb), a fusion protein, and a proprietary bispecific T-cell engager (BiTE) molecule.
We controlled for these critical parameters based on earlier studies establishing their importance:
• material of construction (MoC) — 316L stainless steel
• distance between a coupon and observer (1 m)
• lighting at different intensities and angles of observation
• surface concentration of product.
Analysts spotted precleaned 316L stainless-steel coupons (2 × 2 in2) with
1 mL of mAb, fusion-protein, or BiTE product, monitoring the surface-area coverage of the spotted volume. Surface concentrations ranged 0.25–2.0 µg/cm2. When necessary, the range was extended to 4 µg/cm2. After drying overnight under ambient conditions, spotted coupons were inspected for visible residue from a meter away at three viewing angles (0°, 25°, and 45°).
Native-protein samples had VRLs <4 µg/cm2 across all modalities, and degraded-sample VRLs were <2 µg/cm2. Results were product specific rather than modality dependent, and they were independent of whether a sample contained degraded or native protein.
We used those results to evaluate whether a VRL approach is a viable option for replacing TOC-swab acceptance limits of <0.7 ppm for native protein and <0.5 ppm for degraded protein. Because proteins typically degrade during cleaning processes, VRLs could be used in place of swab acceptance limits as low as 0.5 ppm.
Materials and Equipment
We used the following materials and instruments to conduct our study:
• flat 316L stainless-steel coupons, 2 × 2 in2 (25.8064 cm2) in size
• DS samples (see Table 2 for API and total DS concentrations for each product)
• deionized (DI) water and water for injection (WFI), NaOH, H3PO4, tris base
• 5% (v/v) CIP 100 alkaline detergent (Steris)
• OLS200 shaker bath (Grant Scientific)
• Maglite flashlight
• tape measurer
• autopipette
• illuminance meter
• pH meter
• buffer solutions (pH 4, 7, and 10).
Execution of Experiments
Figure 1 shows the general sequence of our VRL studies. Each step is detailed below.
Figure 1: Sequences of visible residue limit (VRL) testing for samples with degraded/ inactivated and native drug substance (DS).
Calculation of Total Concentrations for Native and Degraded DS Samples: To calculate total DS concentrations, we converted given millimolar values and percentages to mg/mL values using these equations:
Equation 1:
x mM (molec. weight 1000) = mg/mL
Equation 2:
x%(10) = mg/mL
Equation 3:
(mg/mL) + (mg/mL)+ (mg/mL) +... = Total Concentration (mg/mL)
Tables 1 and 2 provide calculations for each DS solution (the mAb, fusion protein, and BiTE molecule).
Table 1: Concentrations (Conc.) of drug substance (DS) samples used in assessment of the visual residue limit (VRL) cleaningvalidation approach; API = active pharmaceutical ingredient, BiTE = a proprietary bispecific T-cell engager (Amgen), FP = fusion protein, mAb = monoclonal antibody.
Table 2: Concentration (Conc.) of drug substance (DS) samples used in assessment of the visual residue limit (VRL) approach; information about individual excipients and molecular weights is not disclosed for reasons of intellectual property. (API = active pharmaceutical ingredient, BiTE = a proprietary bispecific T-cell engager (Amgen), FP = fusion protein, mAb = monoclonal antibody).
Cleaning-Sample Preparation Using a Bench-Scale Inactivation Study: We performed a bench-scale study to simulate a manufacturing-scale cleaning process in which biological products (APIs) are exposed to cleaning agents. Our study served to demonstrate that API was degraded and inactivated during cleaning cycles used at our company’s manufacturing sites. We designed bench-scale cleaning conditions to represent full-scale, real-world cleaning conditions that would be least conducive to API inactivation (worst-case conditions) (Figure 2).
Figure 2: Bench-scale inactivation/degradation study.
A known amount of API was spotted onto each stainless-steel coupon for a specified dirty hold time (DHT), with material stored at a known humidity level. Small-scale DHTs were meant to simulate real-world DHTs, which represent the time between the end of equipment use (e.g., equipment draining) and the start of cleaning.
A reciprocal shaker bath was charged with 15 L of DI water, heated to 70 ± 2 °C (a facility-specific temperature), and set to a speed of 10 cm/s. API-spotted coupons were placed in a sterile vial/bottle with soiled sides facing up. The alkali-wash step was simulated by adding a calculated amount of alkali to reflect full-scale pH conditions. Then, the container was heated to 70 ± 2 °C and placed in the shaker bath at 70 °C for a specified time. Table 3 lists the cleaning-cycle conditions that we applied, and Table 4 provides the resulting data.
Table 3: Cleaning-cycle conditions (Conc. = concentration, WFI = water for injection).
Table 4: Data from laboratory-scale inactivation study; ALK = alkaline, Conc. = concentration, NTZ = neutralization, RE = the ratio of water and caustic solution to drug substance, Vol. = volume.
*RE: R = (Vc + Vw) (Vp), in which Vc is the volume of caustic solution (mL), Vw is the volume of water (mL), and Vp is the volume of product (drug substance).
Calculation of Target Solution Concentrations: After neutralization, each sample had a unique product concentration. To achieve target mass densities (MDs) of 2, 1, 0.5, and 0.25 µg/cm2 within a surface area of 8 cm2, we needed solutions with target concentrations of 16, 8, 4, and 2 µg/mL2:
Equation 4:
(SA x MD) (Vc) = Cs
Therein, SA is the surface area (cm2) covered by the 1-mL spotting volume, which we measured to be ≈8 cm2. Target MDs were 2, 1, 0.5, and 0.25 µg/cm2 for all degraded- and native-protein solutions. The variable Vc represents the volume spotted (1 mL for all samples), and Cs is the concentration of solution required to achieve target MD. Table 6 lists target concentrations for all solutions.
Table 6: Target concentrations (Cs) for degraded- and native-protein solutions; BiTE = proprietary bispecific T-cell engager, FP = fusion protein, mAb = monoclonal antibody, MD = mass density (concentration on coupon surface), SA = surface area covered by 1-mL spot, Vc = volume spotted on coupon.
Dilution of Degraded and Native Protein Solution: Following calculation of targeted solution concentrations, analysts used the following equation to determine requisite dilution conditions:
Equation 5:
C2 = (C1 V1) (V2)
Therein, C2 denotes the target concentration, C1 is the initial concentration of inactivated/degraded solution (µg/mL), and V1 is the initial volume (mL). Dilutions were prepared from inactivated/degraded and native-protein DS samples using WFI
(Tables 7 and 8, respectively).
Table 7: Dilution of degraded-protein solutions from initial (C1) to target concentrations (C2); BiTE = proprietary bispecific T-cell engager, FP = fusion protein, mAb = monoclonal antibody, V1 = sample volume, V2 = target volume.
Table 8: Dilution of native-protein solutions from initial (C1) to target concentrations (C2); BiTE = proprietary bispecific T-cell engager, FP = fusion protein, mAb = monoclonal antibody, V1 = sample volume, V2 = target volume.
Surface-Coverage Study: Several 1-mL samples with product concentrations between 2 and 16 µg/mL were spotted onto each coupon. Analysts confirmed that coverage exceeded 8 cm2 on each 2 × 2 in (25.8064 cm2) coupon. The 8-cm2 value served as a worst-case assumption for calculation because higher coverage values would result in low VRL values.
Coupon Precleaning: Flat 25-cm2 316L stainless-steel coupons were precleaned in 5% (v/v) CIP 100 alkaline detergent at 75 5 °C for a minimum of 10 minutes in a reciprocal shaker bath according to an internal procedure. Following alkaline wash, coupons were rinsed in DI water at room temperature for at least three minutes. Cleaning and rinsing steps were performed in a reciprocal shaker bath at an average linear speed of 10 cm/sec. Following precleaning, clean coupons dried overnight under ambient conditions.
VRL Experiment Setup: During VRL experiments, we controlled for and monitored light intensity, distance between an observer and stainless-steel coupon, and angle of observation (Figure 3). Each parameter was assessed in a wider range than would occur in standard manufacturing conditions to provide worst-case conditions.
Figure 3: Setup of observation distance (X and Y) and angle (θ).
Viewing Station Setup: Analysts prepared the coupon viewing station by placing white cloth on a tray and wrapping a block in white cloth on which to lean the coupons. Then, scientists calibrated a light meter and oriented the light-meter receptor node vertically to match the orientation at which coupons were to be viewed. Lighting shields were adjusted to achieve desired lighting intensities of 250, 700, and 1800 lx. Handheld flashlights emitting white light of 1800 lx were used as needed. Analysts ensured that coupons were positioned to establish 0°, 25°, and 45° observation angles for horizontal and vertical distances as shown in Table 5. Therein, X represents the horizontal distance in inches from the bottom of a coupon to an analyst’s eye, and Y is the vertical distance from the bottom of a coupon to an analyst’s eye. The hypotenuse in this case is held constant at 1 m (39.4 in).
Table 5: Angles and distances used during analysis.
Visual Inspection: After spotted coupons dried, analysts inspected each one visually and recorded results. Increasing the distance between an analyst and coupon lowers the likelihood of confirming soiling visually. The 1-m distance used in our study represents a worst-case scenario compared with visual inspection in a real-world manufacturing environment. Noting whether soiling was visible (V) or not visible (NV), analysts inspected all inactivated/degraded and native DS samples at each lighting intensity and angle of observation.
Looking Ahead
In BPI’s next issue, we will present results from our experiments, including discussion of the influences of lighting intensity and observation angle on coupon evaluation. Using those data,
we will make a case for VRL’s viability as an alternative to using TOC-swab testing with acceptance limits of
<0.7 ppm for native protein and <0.5 ppm for degraded protein.
References
1 ASTM E3263-20. Standard Practice for Qualification of Visual Inspection of Pharmaceutical Manufacturing Equipment and Medical Devices for Residues. ASTM International: West Conshohocken, PA, 2020.
2 ICH Q7. Good Manufacturing Practice Guide for Active Pharmaceutical Ingredients (Step 4 Version). International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use: Geneva, Switzerland, 2000; https://database.ich.org/sites/default/files/Q7%20Guideline.pdf.
3 ICH Q9(R1). Quality Risk Management (Step 4 version). International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use: Geneva, Switzerland, 2023; https://database.ich.org/sites/default/files/ICH_Q9%28R1%29_Guideline_Step4_2023_0126_0.pdf.
4 21 CFR 211.67. Equipment Cleaning and Maintenance. US Food and Drug Administration: Silver Spring, MD, 2023; https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?CFRPart=211&showFR=1&subpartNode=21:4.0.1.1.11.4.
5 EMA/CHMP/CVMP/SWP. Guideline on Setting Health Based Exposure Limits for Use in Risk Identification in the Manufacture of Different Medicinal Products in Shared Facilities. European Medicines Agency: London, UK, 20 November 2014; https://www.ema.europa.eu/en/documents/scientific-guideline/guideline-setting-health-based-exposure-limits-use-risk-identification-manufacture-different_en.pdf.
6 PI 046-1. Guideline on Setting Health Based Exposure Limits for Use in Risk Identification in the Manufacture of Different Medicinal Products in Shared Facilities. Pharmaceutical Inspection Convention and Pharmaceutical Inspection Co-Operation Scheme: Geneva, Switzerland, 1 July 2018; https://www.picscheme.org/layout/document.php?id=1412.
7 EMA/CHMP/CVMP/SWP. Questions and Answers on Implementation of Risk-Based Prevention of Cross-Contamination in Production and “Guideline on Setting Health-Based Exposure Limits for Use in Risk Identification in the Manufacture of Different Medicinal Products in Shared Facilities.” European Medicines Agency: London, UK, 19 April 2018; https://www.ema.europa.eu/documents/other/questions-answers-implementation-risk-based-prevention-cross-contamination-production-guideline_en.pdf.
8 PI 053-1. Questions and Answers on Implementation of Risk-Based Prevention of Cross-Contamination in Production and “Guideline on Setting Health-Based Exposure Limits for Use in Risk Identification in the Manufacture of Different Medicinal Products in Shared Facilities.” Pharmaceutical Inspection Convention and Pharmaceutical Inspection Co-Operation Scheme: Geneva, Switzerland, 2010; https://picscheme.org/layout/document.php?id=2047.
9 Desai P, Walsh A. Validation of Visual Inspection as an Analytical Method for Cleaning Validation. Pharm. Online 11 September 2017; https://www.pharmaceuticalonline.com/doc/validation-of-visual-inspection-as-an-analytical-method-for-cleaning-validation-0001.
10 Ovais M. Statistically Justifiable Visible Residue Limits. Pharm. Technol. 34(3) 2010; https://www.pharmtech.com/view/statistically-justifiable-visible-residue-limits.
11 Fletcher DI. Determination of Surface Visible Residue Limits on Pharmaceutical Plant Equipment. Pharm. Technol. 37(2) 2013: 48–51; https://www.pharmtech.com/view/determination-surface-visible-residue-limits-pharmaceutical-plant-equipment.
12 Forsyth RJ, Van Nostrand V. Visible Residue Limit for Cleaning Validation and Its Potential Application to a Pharmaceutical Research Facility. Pharm. Technol. 28(10) 2004: 58–72; https://www.pharmtech.com/view/application-visible-residue-limit-cleaning-validation-richard-j-forsyth-and-vincent-van-nostrand.
13 Forsyth RJ, Van Nostrand V. Using Visible Residue Limits for Introducing New Compounds into a Pharmaceutical Research Facility. Pharm. Technol. 29(4) 2005: 134–140; https://www.pharmtech.com/view/using-visible-residue-limits-introducing-new-compounds-pharmaceutical-research-facility.
14 Fourman GL, Mullen MV. Determining Cleaning Validation Acceptance Limits for Pharmaceutical Manufacturing Operations. Pharm. Technol. 17(4) 1993: 54–60; https://www.pharmtech.com/view/setting-cleaning-validation-acceptance-limits-topical-formulations.
Corresponding author Ram Kouda, PhD, is a senior principal scientist in process development at Amgen, 1 Amgen Center Drive, Thousand Oaks, CA 91320; [email protected]. Syeda Tabassum is a scientist in process development, Travis Saks-Rudd is a CW-MCS manufacturing associate, and David Dolan, PhD, is scientific associate director of global environmental health and safety (EH&S) at Amgen. Kurt Schnittgrund is a consultant and validation engineer at Azzur Group.
You May Also Like
