Oral chronic Graft-versus-host disease : clinical and pathological staging

Abstract

Allogenic hematopoietic cell transplantation (HCT) is a curative treatment for many patients with immune- hematopoietic disorders, mainly hematopoietic cancers as leukaemia. Chronic Graft-versus-Host Disease (cGVHD) is a major long-term complication, associated with mortality and morbidity following allogenic HCT. Oral cGVHD is common and might manifest as mucosal lichenoid manifestations (om-cGVHD) or with dysfunctional salivary glands (sg-cGVHD). Alloreactive T-cells respond to recipient tissues with pathological reactions of acute inflammation, progressing with chronic inflammation and dysregulated immunity, and subsequent aberrant fibrotic healing.

This thesis aimed to investigate diagnostic criteria for oral cGVHD using histopathological, clinical, and immune cell characterisation. A retrospective cohort of 95 HCT-patients and 303 oral biopsies were analysed, including 15 healthy controls. Oral mucosal biopsies with and without minor salivary glands (MSG) were retrieved from Stockholm Medical Biobank. Associated clinical information was gathered from the clinical charts and HCT register data. We applied histological (Haematoxylin and Eosin, Periodic acid Schiff, van Gieson), and immunohistochemical (IHC) staining (CD4 T-helper cells, CD8 T-cytotoxic cells, CD68 macrophages, CD1a Langerhans cells (LCs), CD19 and CD20 B-cells, and CD5 T-/B-cells). Quantitative IHC was performed using CellProfiler image analysis software.

In papers I and IV, oral mucosal-, and MSG histopathology were analysed in biopsies prior and post HCT, with and without cGVHD. We used the National Institutes of Health pathology criteria and formalised grading modules to assess pathology scores and grades (NIH cGVHD grading). The oral mucosa was observed with minimal criteria of lichenoid interface inflammation with exocytosis, liquefaction degeneration and apoptosis. Basal membrane alterations were the most specific criteria found. Features detected in the MSG were peri-ductal and acinar inflammation and exocytosis, destruction and fibrosis. We developed severity grades (G)0-IV and verified pathology diagnostics of “possible (GII)” and “likely (GIII-GIV)” cGVHD. In paper IV, we also employed the Greenspan composite MSG grading scheme, which was found with a strong correlation to the NIH cGVHD MSG grading.

IHC quantification was performed following established pipelines in CellProfiler, as described in paper II. The methodology was compared to manual counting, with a perfect concordance in detection of positive stained cells, as well as for positive stained regions. The benefit of CellProfiler is to perform standardised and repeatable quantification in a time saving manner.

Oral mucosal immune profiles were investigated in paper III. CD4 infiltration was associated with mild and distinctive om-cGVHD but were found with frequent stable levels over time. CD8 was elevated in clinical and pathologically severe om-cGVHD, particularly during cGVHD onset and progression. Immunolocalisation of CD68 revealed significant staining in various clinical groups, particularly at onset, but the association with severity was interesting especially during late stages of disease. CD1a LCs were significantly reduced in pathological GII at onset and during progression, but otherwise non-significant compared to healthy. CD19 and CD20 were rarely observed.

In paper IV, we quantified the immune profiles in the MSG and found an altered pathology with significant increase of CD4- and CD8-cells. However, levels of B-cells and LCs were considerably low. The association between oral mucosal and MSG immunopathology, was investigated on the whole cohort and with respect to cGVHD duration. Overall, a moderate correlation was observed for pathology scores, CD4 and CD8 infiltrate. Interestingly, at the time of cGVHD onset, the correlation between the oral mucosal disease and MSG was stronger but with progression no further association was found.

In conclusion, om- and sg-cGVHD are two heterogenous complications that display associated immune-pathology profiles during cGVHD onset, but progression appears to be tissue- dependent. We developed histopathological grading modules to facilitate severity diagnostics, which were significantly associated with CD4, CD8 and CD68 immunostaining. om-cGVHD clinical and pathological characterisation was found associated to changes in the immune profile. CD8 was found to drive the severe disease reaction during onset and progression but diminished over time. However, long duration of the disease correlated with elevated CD68 and persistent CD4 cells. This highlights the need for improved clinical and pathological characterisation in combination with biological disease classification.