Programme Hospitalier Recherche Clinique, Institut Pasteur, Inserm, French Public Health Agency.

ObjectiveThe objective of this study was to analyze the specificity of detecting liver tumor cell dissemination by alpha-fetoprotein (AFP) mRNA in peripheral blood. Summary Background DataAlpha-fetoprotein mRNA has been used for the detection of circulating micrometastatic tumor foci of hepatocellular carcinoma (HCC); however, the interpretation of the results has been equivocal. MethodsSixty-four consecutive patients with malignant HCC (n = 20), liver metastases (n = 27), or nonmalignant (n = 17) liver diseases undergoing partial or total hepatectomy and orthotopic liver transplantation were included in this prospective study from January to July 1995. Peripheral blood samples were obtained before surgery, during surgery, and after surgery (range, 6-15 months). Total mRNA was extracted from nucleated cells, and cDNA synthesis and polymerase chain reaction amplification (nested polymerase chain reaction in one tube) were performed with specific AFP primers. ResultsPreoperative AFP mRNA was detected in 20 patients (17%), of which 5 of 20 had HCC. lntraoperative assessment showed positive AFP mRNA values in a total of 34 patients (53%) with various causes, of which 8 of 20 (40%) had HCC, 17 of 27 (63%) had other malignancies, and 9 of 17 (53%) had nonmalignant diseases. Recurrent tumor in patients with HCC occurred in four cases after surgery (range, 6-15 months) and did not correlate with AFP mRNA positivity before surgery, during surgery, or after surgery. ConclusionsAlpha-fetoprotein mRNA in peripheral blood is not a specific marker of circulating micrometastases from HCC, especially in the context of surgical treatment of HCC. 43

Genetic instability has been detected in many types of cancers but poorly investigated in hepatocellular carcinoma (HCC). We have studied the incidence of microsatellite instability (MI) at eight highly polymorphic microsatellite markers and the poly A tract BAT26 and tested for mutations at two sites of repetitive sequence (poly-A nucleotides 709 ± 718 and GT repeatnucleotides 1931 ± 1936) in the Transforming Growth Factor b (TGFb) type II receptor (RII) gene, in a group of 46 European HCCs and the surrounding nontumour tissue. This analysis showed that 63% of HCCs exhibit MI in at least one chromosome locus and 41% in two or more loci. No mutations of the TGFbRII gene were found in the MI positive tumours. No correlation was found with clinicopathological characteristics of the tumours such as cirrhosis, etiology, number of nodules, Edmondson's grade and vascular invasion. However, in patients who had a rearranged D16S402 microsatellite in their tumour, the recurrent disease and the number of nodules were signi®cantly higher than in the others (P50.005 and P50.02, respectively). We propose to consider D16S402 rearrangement in HCC as a prognostic factor to identify patients presenting a higher risk of recurrence.

Summary p53 alterations are considered to be predictive of poor prognosis in hepatocellular carcinoma (HCC) and may induce a humoral response. Anti-p53 serum antibodies were assessed by enzyme-linked immunosorbent assay (ELISA) using purified recombinant human p53 on 130 European HCC patients before treatment and during the clinical course of the disease. p53 immunohistochemistry was performed on tumours from the 52 patients who underwent surgery, and DNA sequencing analysis was initiated when circulating anti-p53 antibodies were detected. Nine (7%) HCC patients had anti-p53 serum antibodies before treatment. During a mean period of 30 months of follow-up, all the negative patients remained negative, even when recurrence was observed. Of the nine positive patients, eight were still positive 12-30 months after surgery. The presence of anti-p53 serum antibodies was correlated neither with mutation of the p53 gene nor the serum alphafetoprotein levels and clinicopathological characterics of the tumours. However, a greater incidence of vascular invasion and accumulation of p53 protein were observed in the tumours of these patients (P < 0.03 and P < 0.01 respectively) as well as a better survival rate without recurrence (P = 0.05). In conclusion, as was recently shown in pancreatic cancer, anti-p53 serum antibodies may constitute a marker of relative 'good prognosis' in a subgroup of patients exhibiting one or several markers traditionally thought to be of bad prognosis.

Lymphoproliferative disorders (LPDs) are a serious side effect of immunosuppression after liver transplantation, and the introduction on the market of a new immunosuppressive drug has been associated with an increased risk of these disorders. To compare the effect of cyclosporine A (CSA) and FK506 in a clinical setting, the incidence of monoclonal or oligoclonal gammopathies known to often precede the appearance of LPDs was evaluated. A total of 88 adult patients was analyzed, 46 were prospectively randomized to CSA and 42 to FK506 for immunosuppression. None of these patients had gammopathy before transplantation. All the patients were tested for immunoglobulin abnormalities five to nine times during a period of 1 year and then two to four times per year thereafter from December 1990 until March 1997. The same incidence of serum immunoglobulin (Ig) abnormalities was observed in both groups (13%) with a mean delay of appearance of 11.1 ؎ 5.9 versus 7.6 ؎ 3.6 months for CSA and FK506, respectively (P G .05). In each group, the gammopathies were transient in 3 patients and persisted in 2. The class of Ig involved was IgG, and a monoclonal component was documented in 2 patients treated with CSA and in 3 patients with FK506. One patient treated with FK506 developed an LPD localized to the lymph nodes 8 months after the occurrence of serum protein abnormalities. The lymphoproliferative lesions subsequently disappeared with the reduction of immunosuppression. In this study, an immunosuppressive regimen of FK506 has not shown an increased incidence of lymphoproliferation compared with CSA in adult liver transplant patients.