Korelacija ekspresije estrogenskih i progesteronskih receptora, ciklina D1 i Bcl-2 sa relapsom kolorektalnih karcinoma Dukes B i C stadijuma

Abstract

Uvod Prema podacima Svetske zdravstvene organizacije i bazi podataka GLOBOCAN, kolorektalni karcinom (KRK) je treći karcinom po učestalosti kod muškaraca a drugi kod žena, i drugi karcinom po smrtnosti u svetu. Oko 70% obolelih biva dijagnostikovano sa lokalizovanom bolešću i potom radikalno operisano. Nažalost, kod oko 1/3 ovih pacijenata se javlja relaps bolesti, a 73% njih umre. Zbog toga je pojava relapsa jedan od najvažnijih faktora preživljavanja obolelih od KRK. Uprkos opsežnom znanju o faktorima rizika za pojavu relapsa, ona se ne može u potpunosti predvideti. Sve je više dokaza da ekspresije ERα, ERβ i PR mogu biti faktori koji utiču na prognozu KRK. Postoje dokazi da je gubitak ekpresije ERβ u tumorskom tkivu povezan je sa višim stadijumima bolesti i predstavlja loš prognostički znak. Prisustvo ERα je otkriveno u debelom crevu najpre na nivou mRNK, ali su nedavne studije potvrdile i njegovu merljivu imunohistohemijsku ekspresiju i prognostiki potencijal. Istraživanja koja su se bavila ekpresijom i ulogom ciklina D1 u patogenezi i prognozi KRK dala su potpuno kontradiktorne rezultate. Prema nekim autorima. povećana ekspresija ciklina D1 je dobar, a prema drugima loš prognostički znak. S druge strane, neki autori su pokazali da se prognostička vrednost ovog markera razlikuje kod muškaraca i žena. Slično je i sa rezultatima studija koje su se bavile značajem ekspresije Bcl-2 u prognozi KRK-a. Neki autori sugerišu da je visoka ekspresija Bcl-2 povezana sa povoljnom prognozom kod pacijenata sa KRK iz Evrope i Amerike, ali su razlike u ekspresiji bile povezane sa primenom neoadjuvantne terapije. Studije koje su ispitivale ekspresiju gore pomenutih markera na tkivu KRK dale su kontradiktorne rezultate u pogledu velikih varijacija nivoa ekspresije i korelacije sa kliničkim ishodom. Cilj naše studije je bio da ispita nivoe ekspresije ERα, ERβ, PR, ciklina D1 i Bcl-2 na tkivu KRK u "visoko rizičnom" Dukes B i Dukes C stadijumu i utvrdi da li postoji njihova korelacija sa ranim relapsom KRK u periodu od dve godine nakon radikalnog hirurškog lečenja. Materijali i metode Dizajn studije i populacija koja se istražuje Ova retrospektivna studija obuhvatila je 101 pacijenta sa dijagnozom kolorektalnog karcinoma. U studiju su bili uključeni pacijenti sa potvrđenim Dukes C i „visoko rizičnom“ Dukes B stadijumom (stadijum pT4 i/ili N2; slabo diferencirani tumor; perforacija; limfovaskularna invazija; perineuralna invazija; <12 pregledanih limfnih čvorova; pozitivne margine nakon operacije). Svi pacijenti su nakon operativnog lečenja primili adjuvantnu hemiterapiju. Iz studije su isključeni pacijenti lečeni od karcinoma rektuma koji su primili neoadjuvatnu hemio/radio terapiju. Pacijenti su podeljeni u dve grupe. Grupu A su činili oni kojima je relaps bolesti dijagnostikovan u prva 24 meseca od operacije (DFS ≤ 24 meseca), a grupu B oni koji nisu imali relaps bolesti najmanje 48 meseci od operativnog lečenja (DFS ≥ 48 meseci). Lečenje i praćenje pacijenata Nakon radikalnog operativnog lečenja i histopatološke potvrde tumora, pacijenti su lečenje nastavili primenom adjuvantne hemioterapije u skladu sa indikacijama Pravilnika o standardima za primenu citostatika Republičkog fonda za zdravstveno osiguranje Republike Srbije, i praćeni na redovnim kontrolama prema preporukama Nacionalnog vodiča dobre kliničke prakse za dijagnostikovanje i lečenje raka kolona i rektuma. Radikalno hirurško lečenje sprovedeno je na Hirurškoj klinici KBC Zemun, a klinički i dijagnostički podaci dobijeni su iz medicinske evidencije koja je formirana tokom adjuvantnog lečenja i redovnog praćenja pacijenata na Službi medikalne onkologije KBC Zemun u periodu od 1. aprila 2008. do 1. januara 2018. Sprovođenje studije odobrila je lokalna Etička komisije za sprovođenje kliničkih studija KBC Zemun. Patohistološko ispitivanje i evaluacija Uzorci tumorskih su preuzeti su iz arhive Službe kliničke patologije KBC Zemun. Semikvantitativna procena ekspresije svih imunohistohemijskih markera izvršena je na osnovu intenziteta imunohistohemijskog bojenja, uzimajući u obzir broj obojenih struktura tkiva i intenzitet imunohistohemijskog bojenja prema sistemu skorovanja datom odvojeno za pojedine vrste markera. Statistička obrada podataka Deskriptivna statistika izračunata je za osnovne demografske i kliničke karakteristike, kao i ishode lečenja. Kategoričke promenljive su predstavljene kao broj i procenat. Kontinuirana distribucija podataka testirana je matematičkim i grafičkim metodama. Kontinuirane promenljive su predstavljene kao srednja vrednost sa standardnom devijacijom (engl. Standard Deviation, SD) ili medijana od 25-75 percentila, prema distribuciji podataka. Razlike između grupa analizirane su pomoću Student T-testa i/ili Man-Whitney-jevog testa za kontinuirane promenljive i Pirsonovog χ2 testa za kategoričke promenljive. Višestruka logistička regresija korišćena je za analizu prediktora za kolorektalni recidiv. Uključene su demografske i kliničke karakteristike od značaja (p < 0.05) i korišćen je unapred uslovni odabir modela kako bi se izbeglo njegovo precenjivanje. Hosmer–Lemeshow test je korišćen za dobro prilagođavanje modelu logističke regresije. Nagelkerke koeficijent je korišćen kao procenat varijanse u zavisnoj promenljivoj povezanoj sa prediktorskom (nezavisnom) promenljivom. Za sve statističke proračune nivo značajnosti (α) iznosio je 0.05. Za statističku obradu dobijenih rezultata koristili smo programski paket SPSS (verzija 23.0, SPSS Inc., Chicago, IL, USA). Rezultati Ova studija je obuhvatila 101 pacijenta radikalno operisanih od kolorektalnog karcinoma u „visoko rizičnom“ Dukes B i Dukes C stadijumu. Prosečna starost pacijenata bila je 64.3 ± 9.9 (DFS ≤ 24 grupa) i 61.1 ± 9.5 (DFS ≥ 48 grupa); nije bilo statistički značajne razlike među grupama (p = 0.105). Medijana praćenja u grupi A bila je 10.0 (3.0-24.0), a u grupi B 69.3 (48.0-115.0) meseci. Učestalost muškaraca i žena bila je slična među grupama; nije uočena značajnija razlika. Nije bilo značajne razlike u lokalizaciji rektuma između grupa. Leva strana debelog creva bila je češće pogođena u grupi sa DFS ≤ 24 meseci; razlika je bila blizu konvencionalnog nivoa značajnosti (p = 0.054). Dukes C stadijum i uznapredovali T stadijum bili su statistički češći u grupi sa DFS ≤ 24 (p = 0.045; p = 0.037). Nije bilo značajne razlike između grupa u N stadijumu i broju pozitivnih žlezda. Limfovaskularna invazija je češće bila prisutna u grupi sa DFS ≤ 24 meseca; razlika je bila blizu konvencionalnog nivoa značajnosti (p = 0.087). Nije bilo razlike između učestalosti perivaskularne invazije i histološkog gradusa među grupama. Učestalost hipertenzije i dijabetesa bila je slične među grupama. U adjuvantnom režimu 84% pacijenata je primilo Mayo hemioterapijski protokol, 13% De Gramont i 3% njih FOLFOX4 protokol. Najčešći komorbiditeti su bili hipertenzija (oko 2/3 pacijenata) i dijabetes (oko 1/5 pacijenata). ERα nije bio ekprimiran na tumorskim tkivima ni jednog pacijenta. ERβ je bio umerereno eksprimiran kod oko 25% pacijenata, češće u grupi sa DFS ≥ 48 meseci u poređenju sa grupom sa DFS ≤ 24 meseca. ERβ je bio je prisutan kod 21 (28.4%) pacijenata sa tumorima levog kolona, i kod 4 (16.7%) pacijenta sa tumorima desnog kolona, razlika nije bila statistički značajna (p = 0.253). Umerena ekspresija PR bila je prisutna kod 1/5 pacijenata, a jaka ekspresija bila je prisutna samo kod dva pacijenta, bez značajne razlike među grupama A i B (p = 0.145). Bcl-2 nije bio izražen kod više od 2/3 pacijenata; 1/3 pacijenata je imalo umerenu ekspresiju ovog receptora, sa sličnom ekspresijom među grupama (p = 0.566). Ciklin D1 je eksprimiran na svim tumorskim uzorcima pacijenata; 15% (14.9) je imalo slabu, 37% (36.6) umerenu, 30% (29.7) jaku i 19% (18.8) vrlo jaku ekpresiju. Jaka i veoma jaka ekspresija ciklina D1 > 50% bila je statistički češća u grupi DFS ≤ 24 nego u grupi DFS ≥ 48 (p = 0.021). Nivoi hemoglobina bili su slični među grupama. CRP je bio značajno veći u grupi sa DFS ≤ 24 (p < 0.001). Nije bilo razlike u NLR i PLR između grupa. Nivo albumina bio je značajno veći u grupi DFS ≥ 48. Pacijenti sa ekspresijom ciklina D1 > 50% su imali 5.2 a sa odmaklim T stadijumom 11.3 puta veću šansu da imaju relaps u toku prve dve godine od radikalne operacije. Umerena ekspresija ERβ je bila udružena sa 79.2% manjom šansom za pojavu relapsa u ovom periodu. Povišene vrednosti CRP, iznad referentnih, bile su udružene sa 5.9 puta većom šansom za kraće vreme do pojave relapsa. Zaključak Pojava ranog relapsa KRK u prve dve godine nakon radikalnog operativnog lečenja u "visoko rizičnom" Dukes B i Dukes C stadijumu povezana je sa smanjenom ekspresijom ERβ i visokom ekspresijom ciklina D1. Ekspresija ERα, PR i Bcl-2 na tkivu kolorektalnog karcinoma nije u korelaciji sa ranim relapsima. Rezultati naše studije ukazuju su povećana ekspresija ciklina D1 i smanjena ekspresija ERβ pokazatelji loše prognoze obolelih od kolorektalnog karcinoma, posebno kod pacijenata sa uznapredovalim T stadijumom. Stoga bi se ciklin D1 i ERβ mogli smatrati značajnim, nezavisnim prognostičkim faktorima za pacijente obolele od ove bolesti i potencijalnim terapijskim metama. Ključne reči: kolorektalni karcinom, relaps, ERα, ERβ, PR, ciklin D1, Bcl-2

Introduction According to the World Health Organization and the GLOBOCAN database, colorectal cancer (CRC) is the third most common cancer in men and the second in women. It is the second most deadly cancer worldwide. About 70% of the patients undergo surgery with curative intent. Unfortunately, around 1/3 of these patients experience disease recurrence, and approximately 73% of them die. Therefore, CRC recurrence is one of the most important factors influencing patient survival. Despite the extensive knowledge about the risk factors for CRC recurrence, its occurrence cannot be fully predicted. There has been growing evidence that the expression of ERα and ERβ, as well as PR, may be factors that influence CRC prognosis. The loss of ERβ expression in tumor tissue is associated with advanced cancer stages and is a poor prognostic sign. ERα was found in the colon at the mRNA level, but recent studies also confirm its measurable immunohistochemical expression and prognostic potential. Researches that have addressed the role of cyclin D1 in the pathogenesis of CRC and its value as a prognostic marker has yielded entirely contradictory results. According to some authors, increased cyclin D1 expression is a good, and according to others, a poor prognostic sign. On the other hand, some authors showed that this marker's prognostic value differs between men and women. The similar is with the results of studies on the importance of Bcl-2 expression in CRC prognosis. Some of them suggest that high expression of Bcl-2 is associated with favorable prognosis in CRC patients from Europe and America, but differences in expression depended on neoadjuvant therapy. The studies examining the expression of the receptors mentioned above on CRC tissue are contradictory in terms of large variations in expression levels and correlations with the clinical outcome. This study aimed to examine the expression levels of ERα, ERβ, PR, Cyclin D1 and Bcl2 on the tissue of CRC in "high risk" Dukes’ B and Dukes’ C stages and determine whether the expression correlates with early CRC recurrence within two years after surgery. Material and methods Study design and study population This retrospective study included 101 patients diagnosed with colorectal cancer in Dukes’ C and "high-risk" Dukes’ B stages (stage pT4 and/or N2; a poorly differentiated tumor; perforation; lymphovascular invasion; perineural invasion; <12 examined lymph nodes; positive margins after surgery). All patients received adjuvant chemotherapy after surgery. Patients treated for rectal cancer who received neoadjuvant chemo/radiotherapy were excluded from the study. Patients were divided into two groups. Group A consisted of those diagnosed with CRC relapse in the first 24 months from surgery (DFS ≤ 24 months), and group B consisted of those who did not relapse for at least 48 months from surgical treatment (DFS ≥ 48 months). Treatment and follow up After radical surgical treatment and histopathological confirmation of tumor, all patients were treated with adjuvant chemotherapy according to the indications of the Ordinance on Standards for the Cytostatic Use of the Republic Health Insurance Fund of the Republic of Serbia. Follow-up was performed according to the recommendations of the National Guidelines for Treatment and Follow-up of Colorectal Cancer Patients. Surgical treatment was carried out at the Surgery Clinic of the Clinical Hospital Center Zemun. The additional adjuvant therapy and follow-up were conducted at the Department of Medical Oncology of the Clinical Hospital Center Zemun from April 1st, 2008. to January 1st, 2018. The study received ethical approval from the Local Research Ethics Committee of the University Clinical Hospital Center of Zemun. Immunohistochemical staining and evaluation Tumor samples were obtained from the archive of the Department of Clinical Pathology of Clinical Hospital Center Zemun. Semiquantitative assessment of all immunohistochemical markers' expression was performed based on the intensity of immunohistochemical staining, considering the number of stained tissue structures and the intensity of immunohistochemical staining according to the assessment system given separately for certain types of markers. Statistical analysis Descriptive statistics were calculated for the baseline demographic and clinical features, as well as treatment outcomes. Categorical variables were presented as numbers and percentages. Continuous data distribution was tested with mathematical and graphical methods. According to data distribution, continuous variables were presented as mean with standard deviation (SD) or median with 25-75 the percentile. Differences between groups were analysed using Student's t-test (or Mann Whitney) for continuous variables and Pearson's Chi-squared test for categorical variables. Multiple logistic regression was used for the analysis of predictors for colorectal recurrence. Significant demographic and clinical features (p < 0.05) were included, and forward conditional model selection was used to avoid model overestimation. Hosmer– Lemeshow test was used for goodness of fit for the logistic regression model. Nagelkerke coefficient was used as a proportion of variance in the dependent variable associated with the predictor (independent) variable. For all statistical calculations, the significance level (α) was 0.05. For statistical processing of the obtained results, we used the SPSS software package (version 23.0, SPSS Inc., Chicago, IL). Results This study included 101 patients with "high risk" Dukes’ B and Dukes’ C stages of colorectal carcinoma. Mean age was 64.3±9.9 (DFS ≤ 24 group) and 61.1±9.5 (DFS ≥ 48 group) without significant difference between groups (p = 0.105). Median follow up in DFS ≤ 24 group was 10.0 (3.0-24.0) and in DFS ≥ 48 group was 69.3 (48.0-115.0) months. The frequency of males and females was similar between groups; no significant difference was observed. There was no significant difference in rectal localization between groups. The left side of the colon was more often affected in DFS ≤ 24 group; the difference was close to the conventional level of significance (p=0.054). Dukes’ stage C and advanced T stage were statistically more frequent in DFS ≤ 24 group (p = 0.045; p = 0.037). There was no significant difference in the N stage and number of positive glands between groups. Lymphovascular invasion was more often present in DFS ≤ 24 group; the difference was close to the conventional level of significance (p = 0.087). There was no difference between perivascular invasion and histological grade between groups. The frequency of hypertension and diabetes were similar between groups. In the adjuvant regimen, 84% of patients received the Mayo chemotherapy protocol, 13% the De Gramont protocol, and 3% of them the FOLFOX4 protocol. The most common comorbidities were hypertension (about 2/3 of patients) and diabetes (about 1/5 of patients). Overall, ERα was not expressed in all patients. ERβ moderate expression was present in 25% of all patients, more often in DFS ≥ 48 group when compare to DFS ≤ 24 group (p = 0.001). Positive ERβ was present in 21 (28.4%) patients with left colon tumors, and in 4 (16.7%) with the right side of the colon, the difference was not significant (p = 0.253). Moderate PR expression was present in 1/5 of patients, and strong expression was present only in two patients, without significant difference between groups (p = 0.145). Bcl-2 was not expressed in more than 2/3 of the patients; 1/3 of the patients had the moderate expression of this receptor, with a similar expression between groups (p = 0.566). Cyclin D1 was expressed in the whole sample of patients; 15% (14.9) had a weak, 37% (36.6) moderate, 30 % (29.7) strong, and 19% (18.8) very strong expression. Strong and very strong expression >50% was statistically more often in DFS ≤ 24 group than in the DFS ≥ 48 group (p = 0.021). Levels of hemoglobin were similar between groups. CRP was significantly higher in the DFS ≤ 24 group (p < 0.001). There was no difference in NLR and PLR between groups. The level of albumins was significantly higher in the DFS ≥ 48 group. Patients with an expression of cyclin D1 > 50% had 5.2, and an advanced T stadium had 11, 3 times higher odds of having a DFS ≤ 24 months. Moderate expression of ERβ was joined with 79.2 % smaller odds for DFS ≤ 24 months. Higher levels of CRP (above referent range) were joined with 5,9 higher odds for DFS ≤ 24 months. Conclusion The occurrence of early relapse of CRC in the first two years after radical surgical treatment in the “high-risk” Dukes’ B and Dukes’ C stages is associated with decreased ERβ expression and high cyclin D1 expression. The expression of ERα, PR, and Bcl-2 on colorectal cancer tissue is not correlated with early recurrences. Our findings indicate the increased expression of cyclin D1 and reduced expression of ERβ are indicators of poor prognosis in CRC, especially in patients with advanced T stage. Therefore, cyclin D1 and ERβ may be considered significant, independent prognostic factors in CRC patients and possible therapeutic targets.