SLC12A2 as a potential histological marker of ulcerative colitis associated colorectal dysplasia

Introduction: Patients suffering from ulcerative colitis (UC) are at increased risk of developing dysplasia
(DAI) and colorectal cancer (CAC). Differentiating DAI from inflammation remains difficult for both
endoscopists and anatomopathologists due to macro and microscopic features shared by these
lesions.
Aim: The aim of our work was to confirm, by histological evaluation, a potential proteomic biomarker
discriminating early DAI lesions from chronic inflamed and normal tissues in UC.
Methods: We included 15 paired tissues from UC patients (n=5) presenting low-grade DAI. Epithelial
cells were isolated by laser capture microdissection and analyzed by label-free proteomics. We
selected one protein differentially distributed between DAI, inflamed (I) and normal (N) tissues for
confirmation by immunochemistry (IHC). IHC characterization was performed using both the staining
intensity score (0 to 4) and the staining pattern: “gradient” (staining intensity increasing from the
epithelium lumen to the bottom of the crypts) or “no gradient” (homogenous staining). UC patients
with DAI (n=28), dysplastic lesion in non-inflammatory colon (DSp) (n=9), CAC (n=14) and at high risk
of CAC (>10 years of UC duration) but free of dysplasia or cancer (n=23) were included. We further
studied this potential marker tissue distribution in the mouse model of CAC (AOM/DSS treated mice)
to trace its presentation at different evolution stages and assessed low (n=51), high-grade DAI (n=35)
and CAC (n=38), as well as relevant paired control tissues. This potential tissue marker was finally
evaluated in sporadic precancerous colorectal lesions of UC-free patients with low (n=19) and highgrade (n=16) adenomas and cancerous lesions (CRC): pT1 to pT4 (n=82) and compared to paired
normal tissues when available.
Results: Proteomics identified 1070 proteins among which 19 showed a differential distribution
between DAI and I or N. The sodium chloride co-transporter SLC12A2 was only identified in DAI.
SLC12A2 IHC “no gradient” staining pattern was associated to DAI and DSp compared to I or N (with p
<0.0001 and 0.0002 respectively). The IHC score was also higher for DAI, DSp and CAC compared to
paired I and N (p<0.0001 and 0.0084 respectively). These results were confirmed from low-grade
dysplasia to more advanced lesions in the AOM/DSS mice model. The “no gradient” pattern was also
significantly associated to low and high-grade adenomas, and CRC of UC-free patients compared to
normal control tissues. The sensitivity and specificity of SLC12A2 histological pattern reached 89% and
95% for DAI versusI; 90% and 93% for CAC and/or DAI versus I. In addition, the sensitivity and specificity
reached 99% and 87% for all precancerous and cancerous lesions (DAI, DSp, CAC and CRC) versus N
and I (including also non-progressing UC patients).
Conclusions: A specific histological pattern for SLC12A2 is associated to precancerous and cancerous
colorectal lesions, and is able to be discriminate these lesions from inflammation and normal tissue in
UC. The continuous upregulation of SLC12A2 in advanced colorectal lesionsin the CAC mice model also
suggests a role of this protein in the pathophysiology of inflammation-associated colon neoplasia.