Accelerated instability testing reveals quantitative mass spectrometry overcomes specimen storage limitations associated with PD-L1 immunohistochemistry.
Author(s): Haragan A, Liebler DC, Das DM, Soper MD, Morrison RD, Slebos RJC, Ackermann BL, Fill JA, Schade AE, Gosney JR, Gruver AM
Publication: Lab Invest, 2020, Vol. 100, Page 874-886
PubMed ID: 31896815 PubMed Review Paper? No
Purpose of Paper
The purpose of this paper was to assess the effects of storing unstained slides under various conditions (different durations, temperatures, humidity, and oxygen content) on programmed-death-ligand-1 (PD-L1) immunohistochemistry (IHC) staining of gastric tumor, non-small cell lung carcinoma (NSCLC), tonsil, and placental sections. The effects of storage under extreme conditions on quantification of PD-L1 by targeted mass spectrometry (MS) analysis and proteome-wide oxidation were also studied.
Conclusion of Paper
Gastric tumor sections stored for 4.5 or 24 months under normal conditions had significantly lower PD-L1 positive pixel scores than unstored sections but the magnitude of the decline depended on the antibody used. Similarly, storage of NSCLC under accelerated conditions (37°C, 100% oxygen, and 80% humidity) resulted in a progressive loss of PD-L1 immunoreactivity for three of four antibodies with the fourth producing low staining prior to storage. Storage of placenta and tonsil sections under accelerated conditions (37°C, 100% oxygen, and 80% humidity) also resulted in a progressive loss of PD-L1 and pan-cytokeratin (CK) immunoreactivity that was not observed during storage under normal conditions. The loss in PD-L1 immunoreactivity was found to accelerate with increasing temperature and/or humidity but was not affected by oxygen content. Further evaluation using tonsil specimens showed that the loss of PD-L1 positivity occurred at a similar rate in crypt epithelia and germinal centers but, because the initial signal was weaker, immune cell PD-L1 loss was observed sooner. Importantly, the initial staining and loss of immunoreactivity as evaluated by combined positive score (CPS) and the tumor proportion score (TPS) were dependent on the antibody used. Targeted MS analysis showed no loss in oxidizable or non-oxidizable PD-L1 peptides in NSCLC sections with storage but more variation in levels, particularly of the oxidizable peptide, were observed. Global MS analysis showed increased oxidation in placenta and tonsil section stored under high humidity for 28 days or 2 years.
Studies
-
Study Purpose
The purpose of this study was to assess the effects of storing unstained slides under various conditions (different durations, temperatures, humidity, and oxygen content) on PD-L1 IHC staining of gastric tumor, non-small cell lung carcinoma (NSCLC), tonsil, and placental sections. The effects of storage under extreme conditions on quantification of PD-L1 by targeted MS analysis and proteome-wide oxidation were also studied. FFPE blocks of 10 non-small cell lung cancer (NSCLC), 35 gastric carcinoma, 6 placenta, and 6 tonsil specimens were procured from commercial sources (no details provided). Serial 4 µm sections were allowed to dry overnight at room temperature and then immediately used for IHC staining or stored for 0, 4.5, or 24 months under normal conditions (gastric sections) or 0, 1, 3, 9, 19, or 28 days under accelerated conditions (NSCLC sections). Normal storage was at 20.1-31.0 °C (average 21.6°C) and 14.4-80.5% humidity (average 46.8%) and, unless otherwise specified, accelerated storage was in a chamber at 37°C, 100% oxygen, and 80% humidity. IHC was performed using four different antibodies against PD-L1 and cytokeratin. Representative sections were also hematoxylin and eosin stained. Staining was assessed by pathologists who calculated the TPS based on the number of PD-L1 positive tumor cells as well as tumor cell/immune cell scoring. PD-L1 and CK staining was also assessed using Aperio ImageScope integrated image analysis which calculated positivity based on the percentage of positive pixels and calculated a combined positive score based on the count of positive tumor and immune cells. PD-L1 levels were also assessed by reverse-phase liquid chromatography targeted MS and the proteome analyzed on unfractionated tryptic digests by reverse-phase liquid chromatography-MS.
Summary of Findings:
Gastric tumor sections stored for 4.5 or 24 months had significantly lower PD-L1 positive pixel scores than unstored sections with both the EIL3N (0.197 versus 0.107; P= 0.05 and 0.197 versus 0.070; P<0.001, respectively) and SP142 (0.128 versus 0.075; P<0.001 and 0.128 versus 0.074; P<0.001, respectively) antibodies. Importantly, the CPS was also higher for the E1L3N antibody than for SP142 antibody (40 versus 30, P<0.05), but there was a clinically relevant decline in CPS with both antibodies (from ≥1% to less than 1%) in many slides after 4.5 months (13% for EIL3N and 20% for SP142) with even more slides showing loss by 24 months (33% for EIL3N and 37% for SP142). While staining with 22C3 and 28-8 antibodies was comparable based on both staining and percent with TPS on Day 0 ≥1% and ≥50%, respectively, the EIL3N antibody had slightly higher average staining but comparable percentage with TPS ≥1% and ≥50%, respectively, and SP142 much lower staining as well as percentage TPS ≥1% and ≥50%, respectively. During storage under accelerated conditions, there was a progressive loss in 22C3, 28-8, and E1L3N immunoreactivity and TPS in NSCLC sections but the already low staining with SP142 did not decrease further. Similarly, storage of placental and tonsil sections under the accelerated conditions resulted in loss of PD-L1 (83% loss, P<0.001 and 85% loss P<0.001, respectively) and pan-CK positivity (59% loss, P=0.05 and 31% loss P<0.05), but no loss in immunoreactivity was observed when sections were stored under normal conditions. Increasing the temperature from 37°C to 60°C accelerated the loss in PD-L1 immunoreactivity such that after 7 days only 8% and 3% of the original immunoreactivity remained in placental and tonsil sections, respectively. Decreasing the temperature in the chamber increased the stability of the immunoreactivity such that by Day 28 only the loss in PD-L1 in tonsil sections was significant (48%, P<0.05). Similarly, reducing the humidity in the chamber from 80% to 45% while retaining the 37°C temperature and 100% oxygen slowed the loss in PD-L1 immunoreactivity such that a loss was only significant in tonsil specimens after 28 weeks (90%, P<0.05) and was nonsignificant in placental sections after 39 weeks. Importantly, the PD-L1 immunoreactivity loss was comparable in sections stored for 1 week at 80% humidity and 39 weeks at 45% humidity. Placing the slides in a container with desiccant inside the chamber at 80% humidity decreased the relative humidity to 30% and thus the loss in PD-L1 immunoreactivity was comparable to that under normal conditions. There was no difference in loss of PD-L1 or pan-CK immunoreactivity when stored in a chamber at 100% oxygen versus 20% oxygen. In tonsil specimens, loss of PD-L1 positivity occurred at a similar rate in crypt epithelia and the germinal centers but, because the initial signal was weaker, immune cell PD-L1 loss was observed sooner.
Targeted MS analysis showed no loss in oxidizable or non-oxidizable PD-L1 peptides in NSCLC sections with storage but more variation in levels, particularly of the oxidizable peptide, were observed. However, global MS analysis showed increased oxidation in placenta and tonsil sections stored for 28 days under high humidity (P=9x 1014) or for 2 years under normal conditions (1 x 10−34 for tonsil and P= 8 × 10−58 for placenta compared to baseline for accelerated specimens).
Biospecimens
Preservative Types
- Formalin
Diagnoses:
- Not specified
- Neoplastic - Carcinoma
Platform:
Analyte Technology Platform Protein Immunohistochemistry Peptide LC-MS or LC-MS/MS Protein LC-MS or LC-MS/MS Pre-analytical Factors:
Classification Pre-analytical Factor Value(s) Biospecimen Acquisition Biospecimen location Tonsil
Placenta
NSCLC
Gastric tumor
Storage Storage duration 0 days
1 day
3 days
9 days
19 days
28 days
0 months
4.5 months
24 months
Storage Storage temperature 20.1—31.0°C
20°C
37°C
60°C
Storage Storage conditions Normal
Accelerated
80% humidity
30% humidity
100% oxygen
20% oxygen
Immunohistochemistry Specific Type of antibody 22C3
28-8
SP142
E1L3N
LC-MS or LC-MS/MS Specific Technology platform Targeted
Global