Host cell proteins and residual host cell DNA are process-related impurities that every sponsor of a monoclonal antibody or recombinant protein has to control, and they come up often in FDA's CMC review of a BLA. Because FDA sets no universal numerical limit, sponsors have to justify their acceptance criteria and show that their assays can detect the impurities their own production cell line actually makes. Gaps in either area can lead to information requests, postmarketing commitments or changes to the lot-release strategy.
The analysis below draws on FDA guidance and published BLA review documents. It covers how reviewers have assessed HCP and residual DNA acceptance criteria, how they have judged whether HCP assays are suitable and whether the antiserum covers the cell line's HCPs, when they have used postmarketing commitments to deal with assay limitations, how they have handled residual DNA testing and clearance validation, and how these expectations apply to biosimilar applications.
FDA evaluation of host cell protein and residual DNA in BLA reviews for monoclonal antibodies and recombinant proteins
Key takeaways
- FDA treats host cell protein (HCP) and residual host cell DNA as process-related impurities controlled through product- and process-specific assays and acceptance criteria. No single numerical limit applies to all therapeutic proteins 8687.
- In the BLA reviews examined, a recurring CMC question is HCP assay suitability, especially whether the anti-HCP antiserum covers the HCPs of the actual production cell line. Coverage is shown by comparing a sensitively stained 2D SDS-PAGE with a western blot probed with the assay antiserum 106107108.
- When coverage is incomplete, FDA has often accepted the existing assay at approval and required a postmarketing commitment (PMC) to develop a more sensitive, process-specific HCP assay. Examples include Keytruda, Opdivo, Unituxin, Tepezza and Gamifant 3537404542.
- For residual DNA, the reviews examined mostly show qPCR results below detection or quantitation limits. Validated clearance can support removing DNA testing from lot release, as FDA accepted for Skyrizi 206.
- For biosimilars, FDA's current position is that HCP and DNA profiles do not need to match the reference product. They are evaluated through process controls and a risk-based assessment, including immunogenicity risk 91157164.
Regulatory framework reviewers apply
The BLA review findings below follow from a small set of guidance documents.
ICH Q6B. Process-related impurities should be minimized through a well-controlled manufacturing process. Acceptance criteria are set individually or collectively as appropriate, based on data from clinical, preclinical and manufacturing-consistency lots. In some circumstances, a selected impurity may not need an acceptance criterion 8687. Q6B generally calls for a sensitive HCP immunoassay that detects a wide range of proteins. The assay's polyclonal antibody is raised against a production-cell preparation that lacks the product-coding gene. Host cell DNA can be measured directly in the product, and demonstrated clearance of HCP and nucleic acids (for example, by laboratory-scale spiking studies) can sometimes remove the need for acceptance criteria 88.
CMC submission guidance for recombinant DNA-derived products and monoclonal antibodies. The guidance asks for purification-validation documentation showing adequate removal of residual host proteins and DNA, where appropriate 89.
Points to Consider in the Manufacture and Testing of Monoclonal Antibody Products for Human Use. Lot-to-lot DNA testing before excipient addition is recommended unless a properly conducted DNA-clearance study substitutes for it. Final-product DNA should be as low as possible. The document suggests no more than 100 pg cellular DNA per dose whenever possible, using a method sensitive to 10 pg. It also calls for tests to detect and quantify HCP 90. The 100 pg/dose value is a 1997 suggested target, not a universal regulatory limit that reviewers apply across products 90.
Immunogenicity Assessment for Therapeutic Protein Products. Host-cell-related impurities, including proteins and nucleic acids, can enhance immune responses. Even trace levels may matter, so assays should be sensitive enough and tailored to the relevant cell substrate 92. The comparability protocols guidance likewise expects assessment of potentially immunogenic impurities such as HCP when a manufacturing change is made to a biological-source product 93.
Biosimilar comparative analytical assessment guidance. HCP and host cell DNA in the proposed biosimilar should be evaluated, but their profiles are not expected to match the reference product. They are generally not part of the comparative analytical assessment. Because HCP immunoassay results depend on the reagents and the cell substrate, the assays should be validated using the proposed product's cell substrate, with orthogonal methods supporting accuracy and sensitivity 91.
HCP assay suitability: the core review issue
What FDA asks applicants to show
Across BLA reviews, FDA considered an HCP ELISA suitable only when the antiserum's coverage of relevant production-cell or process HCPs was demonstrated, together with adequate sensitivity and qualification or validation. Reviewers asked applicants to compare a sensitively stained two-dimensional SDS-PAGE HCP pattern with a western blot probed with the assay antiserum, and to estimate the percentage of potential HCPs the assay recognizes. A similarly sensitive and discriminating method could be used instead. One-dimensional gels alone were considered insufficient 106107108.
Product-specific findings
| Product | Expression context | FDA finding or request |
|---|---|---|
| Yervoy (ipilimumab) | mAb | Requested the HCP assay validation report and evidence that the antibody detects most potential process HCPs, using a sensitive gel/blot comparison or an equivalent method 108111. FDA also asked for clarification of CHO protein as a process-related impurity 2. |
| Adcetris (brentuximab vedotin) | ADC | Requested qualification of the anti-HCP antiserum by sensitive-stain 2D SDS-PAGE compared with western blot, or a similarly sensitive assay 108111. |
| Opdivo (nivolumab) | mAb | Filing with the current assay might be acceptable if clinical-development data supported it. FDA expected a better-coverage assay to be developed and implemented after licensure, and asked that available information on it be included in the BLA 112. |
| Trazimera (trastuzumab-qyyp) | Biosimilar mAb | Reviewers found the ELISA antibodies' overall coverage poor and the evidence insufficient to support the assay. They requested more data on coverage of HCPs specific to the production cell line 110. |
| Inflectra (infliximab-dyyb) | Biosimilar mAb (Sp2/0) | FDA requested assay and antiserum descriptions, qualification and comparability methods, coverage and sensitivity data relative to the original assay, and 2D-gel/western-blot coverage data. FDA warned that commercial kits are usually insensitive to product-specific HCPs 109. |
| Releuko (filgrastim-ayow) | Biosimilar recombinant protein | Changes to commercial-kit antibodies or other critical reagents could alter ELISA performance. FDA requested reagent control and requalification of new kit lots, including 2D SDS-PAGE/western blot coverage. FDA identified an in-house qualified antiserum as the best approach 115. |
| Myalept (metreleptin) | E. coli | The commercial ELISA's claimed 74% antigen-pool coverage was not supported by data. Inspection data suggested inadequate coverage of the actual production cell line. The assay appeared adequate in routine operation, but reviewers were unsure whether it would detect increased HCP after a deviation 113. |
| Tzield (teplizumab-mzwv) | mAb | The existing assay was acceptable for drug substance release. FDA sought product-specific HCP antibodies to ensure complete coverage and required a PMC 114. |
Two themes stand out. First, FDA does not accept generic or commercial-kit ELISAs by default. Applicants must show coverage for the specific cell line and process, and must control critical reagents across kit lots 109115. Second, reviewers separate "adequate for routine release" from "able to detect a process excursion." Myalept is the clearest example: an assay can pass routine lots and still fail to flag an increase in HCP during a deviation 113.
Orthogonal methods
In the BLA review passages examined, FDA's coverage checks relied on gel/blot comparisons or similarly sensitive alternatives. These passages did not show reviewers requesting mass-spectrometry identification of individual HCPs or naming specific high-risk HCPs such as polysorbate-degrading lipases 209210211. The biosimilar guidance does call for orthogonal methods to support HCP immunoassay accuracy and sensitivity 91. The absence of such requests in these passages should not be read as FDA practice; current expectations, including USP <1132> and orthogonal MS characterization of high-risk HCPs, may not be reflected in the review text examined.
Postmarketing commitments on HCP assays and limits
When coverage or sensitivity is not fully established at approval, FDA has repeatedly used PMCs instead of delaying approval. These commitments follow a consistent template:
| Product | Commitment |
|---|---|
| Keytruda (pembrolizumab) | Develop and validate a more sensitive, process-specific HCP assay with broader detection. Implement it for drug substance release and submit the procedure, validation report, proposed acceptance criterion and supporting data 35. Separate PMCs require re-evaluation of drug substance and drug product specifications after 30 commercial-scale lots 36. |
| Opdivo (nivolumab) | A parallel process-specific HCP assay commitment 37, with specification re-evaluation after 30 lots 3839. |
| Unituxin (dinutuximab) | Develop a process-specific assay with better sensitivity and broader coverage. Evaluate antiserum coverage by 2D SDS-PAGE and western blot, and submit the procedure, validation, gel/blot reproductions, coverage analysis and proposed acceptance criterion in a prior-approval supplement 40107. A separate PMC requires specification reassessment 41. |
| Gamifant (emapalumab-lzsg) | Complete a more sensitive HCP assay, test all clinical, PPQ and commercial drug substance lots, then re-evaluate the HCP release acceptance criterion with statistical analysis 42. Specifications are re-evaluated after 12 additional commercial lots 43. |
| Tepezza (teprotumumab-trbw) | Develop, validate and implement a more sensitive, product-specific HCP assay. Submit it as a CBE-30 supplement 45. |
| Ontruzant (trastuzumab-dttb) | Perform a qualification study using production material to confirm the assay detects product-specific HCP in drug substance 44. |
| Tzield (teplizumab-mzwv) | Confirm current HCP assay suitability, defined by HCP coverage 46. |
| Margenza (margetuximab-cmkb) | Include residual HCP in a postmarketing reassessment of drug substance specifications, either tightening or supporting the existing criterion 192. |
| Xiaflex (collagenase C. histolyticum) | Demonstrate the feasibility of an immune-based HCP assay and propose a path forward 34. |
The pattern shows that the HCP acceptance criterion depends on the assay. Gamifant illustrates this most directly: FDA tied re-evaluation of the HCP release limit to retesting historical lots with the improved assay 42. Numerical HCP acceptance criteria (ng/mg or ppm) were generally not stated or were redacted in the review text examined 138191.
Residual host cell DNA: limits, assays and clearance
Residual DNA drew fewer review comments than HCP. Most findings in the reviews concern qPCR results at or below assay limits:
- Kanjinti (trastuzumab-anns): qPCR on 22 drug substance lots and three lots each of US-licensed and EU-approved Herceptin found DNA below the 0.2 pg/mg detection limit in all three products. The data supported similarity for this attribute 50. The 0.2 pg/mg figure is a detection limit, not a release limit 50.
- Mvasi (bevacizumab-awwb): qPCR results were reported as below 1 pg/mg, with two EU-Avastin lots below 2 pg/mg. The reviewer found residual DNA levels similar across the three products 204205.
- Herzuma (trastuzumab-pkrb): qPCR found DNA below the detection limit in every tested biosimilar and reference lot 51.
- Nivestym (filgrastim-aafi): DNA results by qPCR were below LOQ for both the biosimilar and US-licensed Neupogen lots. HCP by ELISA was below or near the 0.5 ng/mL LOQ. The reviewer judged host-cell impurities to be a low product-quality risk 186176162.
- Myxredlin (insulin human) and Rylaze (crisantaspase): Drug substance specifications include a host-cell DNA test (Myxredlin's limit is expressed per 3.5 mg), but the numerical values are redacted 207208.
Clearance-based removal from release testing
Consistent with Q6B and the monoclonal antibody Points to Consider 8890, FDA accepted demonstrated clearance as a basis to remove residual DNA testing from lot release for Skyrizi (risankizumab-rzaa) 206. Pre-submission correspondence shows FDA's conditions for this approach:
- Cyramza (ramucirumab): Clearance of process-related impurities to below detection or quantitation limits could potentially justify removing them from the drug substance specification. FDA reserved its decision until BLA review and asked for the risk assessment, evidence that the small-scale model represents the commercial process, and qualification of the HCP assay 57.
- Ajovy (fremanezumab-vfrm): FDA asked for a risk assessment estimating the process's impurity-removal capability and theoretical residual levels. It said theoretical clearance might support a specification justification 58.
The Cyramza correspondence shows that HCP assay qualification is a prerequisite for clearance claims: FDA cannot rely on clearance data generated with an assay whose coverage has not been established 57.
Biosimilars: risk-based assessment, not matching
FDA's biosimilar reviews increasingly reflect the guidance position that process-related impurities need not match the reference product 91:
- Nyvepria (pegfilgrastim-apgf): FDA considered direct HCP/DNA comparison inappropriate because the manufacturing processes differ. It accepted commercial process controls, with HCP and residual DNA results below the reported limits across roughly 30 lots 157.
- Rezvoglar, Semglee (insulin glargine) and Stoboclo/Osenvelt (denosumab-bmwo): Reviewers took the same approach and assessed HCP and DNA through commercial process controls 158159160161.
- Ziextenzo (pegfilgrastim-bmez): FDA recommended side-by-side assessment of process-related impurities without expecting them to match. Different or higher impurity levels would require a risk assessment and possibly more studies, because of potential effects on safety, purity and potency 164.
- Udenyca (pegfilgrastim-cbqv): 2D gels showed minor HCP differences from Neulasta. The reviewer accepted the HCP risk assessment, which found very little risk, but noted residual uncertainty given Udenyca's higher incidence of immunogenicity. The review did not establish that the HCP differences caused it 126.
Earlier trastuzumab and bevacizumab biosimilars (Kanjinti, Mvasi) did report comparative DNA and HCP data as supportive tier 3 similarity assessments 187188205. Later reviews more often rely on process control instead.
Non-CHO expression systems and immunogenicity
For microbial and yeast-expressed proteins, reviewers linked HCP control directly to immunogenicity:
- Myalept (metreleptin, E. coli): Reviewers noted that HCP can act as an adjuvant, and that neutralizing antibodies to metreleptin have been associated with loss of efficacy or of endogenous leptin activity. Even so, FDA allowed the assay work to continue after approval, citing the controlled process and the seriousness and rarity of the disease 144145.
- Semglee (insulin glargine, yeast): The immunogenicity reviewer found the anti-yeast-HCP antibody assay adequately validated. The proportions of anti-HCP antibody-positive patients were comparable between Semglee and Lantus 146.
- Qwo (collagenase clostridium histolyticum-aaes): FDA's quality risk table classified production-cell HCP as a safety and immunogenicity risk, and production-cell DNA as a safety risk 147.
- Somavert (pegvisomant): Reviewers evaluated the specificity and detection limit of the residual E. coli protein assay 212.
Practical implications for BLA sponsors
Based on these reviews, sponsors should expect FDA to focus on the following:
- Coverage data for the HCP assay. Provide 2D SDS-PAGE/western blot (or an equivalent sensitive method) coverage data for the actual production cell line. A commercial kit's claimed coverage is not enough 108113109.
- Readiness for a PMC on the HCP assay. If a platform or commercial assay has limited coverage, FDA may accept it at approval and require a process-specific assay with re-evaluation of the acceptance criterion after licensure 354245.
- Critical reagent control. Requalify coverage when antiserum or kit lots change 115.
- Clearance claims built on qualified assays and representative scale-down models. These are the basis for removing DNA or HCP from release testing 57206.
- A risk-based justification for differences in biosimilar programs. Do not try to match reference-product HCP and DNA levels; justify any difference by risk, with attention to immunogenicity 91164126.
- Clinically and process-based acceptance criteria. Base limits on clinical and manufacturing-consistency lots, and expect re-evaluation after a defined number of commercial lots 863643.