The value of microsatellite instability in the detection of HNPCC families and of sporadic colorectal cancers with special biological features: An investigation on a series of 100 consecutive cases
Annals of Oncology
D. Furlan 1
M. G. Tibiletti 0 1
M. Taborelli 1
L. Albarello 1
M. Cornaggia 0 1
C. Capella 1
0 Department ofPathology, Ospedale di Circolo , Varese , Italy
1 department of Clinical and Biological Sciences, University ofPavia at Varese
The value of microsatellite instability in the detection of HNPCC families and of sporadic colorectal cancers with special biological features: An investigation on a series of 100 consecutive cases Summary Results: MI was detected in 36 of 100 cancers, 27 of which showed low instability and nine a high instability. The low- and Background: Microsatellite instability (MI) is a biological char- high-instability cases showed similar clinicopathological characteristic of most tumors involved in hereditary non-polyposis acteristics, and significantly positive associations were observed colorectal cancer (HNPCC). This disease appears to be caused between MI and mucinous histological type (P = 0.0001) and by germline mutations in mismatch repair (MMR) genes, MI and peritumoral lymphocytic infiltration (P = 0.01). A which are responsible for repairing single base-pair mis- single HNPCC family was identified in the high-grade MI matches. At least five human genes participate in MMR. MI group, while two families belonged to the Mi-negative group. also occurs in 10%-15% of sporadic colorectal cancers. Be- Conclusions: Our data suggest that MI screening is probcause MI detection has been suggested as an alternative diag- ably not an efficient strategy for identifying HNPCC cases. MI nostic tool for identification of HNPCC families, in this study does, however, appear capable of defining a category of colowe analyzed the MI pattern in 100 consecutive colorectal carci- rectal cancers with favourable prognostic features and should nomas in order to correlate them with the clinicopathologic be investigated at least in all cases of mucinous adenocarcinofeatures and family histories of the patients. mas. Patients and methods: A series of 100 colorectal cancers was evaluated for MI with 10 polymerase chain reaction primer sets. Instability results were compared with family history and Key words: colorectal cancers, histological type, HNPCC, other clinical and biological characteristics. RER phenotype
Hereditary non-polyposis colorectal cancer (HNPCC)
is an autosomal dominant disease characterized by the
early (i.e., in patients below the age of 45-50) occurrence
of colorectal cancers. The disease appears to be due to
germline mutations in the mismatch repair (MMR)
genes which are responsible for repairing single
basepair mismatches. The generally accepted mechanism for
HNPCC tumorigenesis according to Knudson's model
 involves inheritance of a single recessive MMR
mutation with subsequent inactivation of its wild-type allele
counterpart in a somatic cell, which is likely to occur in
highly proliferating tissues . Inactivation of MMR
genes leads to genomic instability, characterized by
either the expansion or contraction of short repeat DNA
sequences [3, 4]. This form of instability, known as
microsatellite instability (MI) or as replication error
(RER), is thought to cause a rapid accumulation of
somatic mutations in different oncogenes and tumor
suppressor genes which play crucial roles in tumor
initiation and progression [5-7]. At least five human genes
participate in MMR. Germline mutations in these genes
have been identified in HNPCC families, namely,
hMSH2, hMLHl, hPMSl, hPMS2 and MSH6/GTBP
[8-12]. More than 90% of mutations in HNPCC
families are found in hMSH2 and hMLHl, while it appears
that the hPMSl, hMSH6 and hPMS2 genes are only
rarely mutated .
It has been demonstrated in verified HNPCC patients
that MI is frequently associated with colorectal
carcinomas [3, 13]. A population-based screening for MMR
gene mutations to identify subjects at risk for HNPCC
is time-consuming and extremely expensive because of
the involvement of multiple genes and the heterogeneity
of mutations in such patients [14, 15]. HNPCC is
currently identified by pedigree reconstruction, using the
so-called 'Amsterdam criteria' . Although Jass et al.
 recently suggested the use of RERs as an alternative
diagnostic tool for this disease, several studies have
demonstrated that the RER-positive phenotype is also
detected in a considerable fraction (about 15%) of
sporadic colorectal cancers and that by itself it is not an
indicator of the presence of familiarity [13,18].
In this study we analyzed RER phenotypes at
multiple loci in 100 consecutive colorectal carcinomas with
the aim of: 1) correlating them with the
clinicopathologic features and family histories of the patients; 2)
comparing their value in identifying HNPCC cases to
that of pedigree reconstruction.
Patients and methods
Patients and samples
We investigated a consecutive series of 100 patients (98 unrelated and
two belonging to the same family) who underwent surgery for colorectal
cancer at the Varese Hospital from November 1995 to November 1996.
In each case the patient's age and tumor site and size were
recorded. The tumors located proximal to the splenic flexure were
classified as right-sided and the tumors distal to the splenicflexureas
left-sided In each case fresh tissue fragments from primary tumor,
contiguous to the samples taken for conventional pathological
investigation, and samples of non-neoplastic mucosa were stored at -80 C
for molecular investigation. The presence of non-malignant
'contaminating' cells in the tumor specimens was evaluated on contiguous
histological sections: in all samples the cut-off value of 50% for the
proportion of malignant cells was used, but about 3/4 of the cells in
most of the samples were malignant.
The tissues were fixed in formalin and embedded in paraffin.
Hematoxylin- and eosin-stained sections were used to categorize tumors
according to the WHO Classification  The adenocarcinomas were
graded into well- (Gl), moderately- (G2) and poorly
(G3)-differentiated. Tumors were classified as mucinous when the amount of mucin
was more than 50% of the examined areas.
In all cases, the following histological parameters were evaluated,
using a two-step scoring system (absent or minimal and present):
pentumoral lymphocytic infiltration, necrosis, stromal fibrosis,
vascular and neural invasion. The mitotic index was expressed as the number
of mitoses per 10 high-power fields (HPF).
The neoplasm extension was evaluated according to the TNM
Genomic DNA was isolated from frozen tissues using QIAamp tissue
kits (Qiagen Germany). In 22 cases three different samples from the
same tumor were used to evaluate tumor heterogeneity.
DNA from paired normal and tumor tissues were PCR-amplified
at 10 markers: D2S123 (2ql6), D2S119 (2pl6), D3S1611 (3p22-p24),
APC, D7S481 (7pl5-pter), D7S517 (7pl5-pter), D13S175 (13qll),
The PCR mixture (50 vl) included 100-200 ng of genomic DNA,
lOmM Tris, 50mM KC1, 1-2 mM MgCl2, 200 uM dNTPs, and 50
pmol of each primer, 2U Taq Polymerase (Perkin-Elmer, Italy).
Samples were denatured at 94 C for five minutes followed by 30 cycles of
denaturation (94 C, one minute), annealing (55 C, one minute),
extension (72 C, one minute), and a final five-minute extension at 72 C
in a Perkin-Elmer Gene Amp Thermal Cycler 480. PCR products were
electrophoresed on 6% denaturing polyacrylamide sequencing gels for
two hours at 1600 V, 40 mA, 50 C. Polyacrylamide gels,fixedon glass
plates with methacryloxypropyltrimethoxysane, were silver-stained
using DNA Silver Staining System (Promega, Italy). Silver-stained gels
were reproduced on EDFfilms(Kodak, USA).
Genetic counselling with psychological support was carried out at the
first clinical follow-up of each patient. Using clinical reports or death
certificates for validation of cancer cases, family medical histories of at
least three generations were established.
For HNPCC identification the so-called 'Amsterdam criteria' were
adopted . The families with one or two missing 'Amsterdam criteria'
were considered suspect for HNPCC .
Associations between variables were tested by Pearson's x2"test o r
Fisher's exact test when appropriate.
MI was detected in 36 of 100 colorectal carcinomas: 27
of them were unstable at 1 or 2 loci, while nine were
unstable at three or more loci. Both microsatellite
expansions (Figure 1) and contractions were identified
independently of the number of altered loci. The loci
more frequently altered were APC (11 cases), D2S123
(10 cases), and D3S1611, D7S481 and DCC (nine cases
each). Tumors showing MI were divided into two groups
according to the number of involved loci. Twenty-seven
carcinomas showed low instability (MI at 1 or 2 loci) and
nine high instability (MI at more than 2 loci). The
unstable loci were different in the two groups, as shown in
Figure 2. The D13S175 locus was exclusively involved in
high-instability cases, while D3S1611 and DCC were
more frequently involved in low-instability cancers. In
the 22 cases in which multiple samples of the same
tumor were analyzed we detected complete concordance
among altered MI loci in all fragments of the same case.
The relationships between MI, clinico-pathological
data and family history are listed in Tables 1 and 2. The
low- and high- instability cases showed similar
clinicopathological characteristics. Among the various
clinicopathologic parameters examined (Table 1) only two, the
mucinous histological type (P - 0.0001) and the
presence of peritumoral lymphocytic infiltration (P - 0.01),
were significantly associated with both low- and
MI absent (64 cases)
Low MIa (27 cases)
High MIb (nine cases)
Total (100 cases)
P-value 64.4 14.7 23/41 19
Grade 1 2
Lost to follow-up
Number of relatives with malignant tumors
of any site
Number of relatives without malignant
Number of relatives with colorectal cancers
Number of relatives without colorectal
Number of relatives with malignant tumors
other than colorectal0
Number of relatives without malignant
tumors other than colorectalc
MI absent (64 cases, Low MIa (27 cases)
High MIb (nine cases) Total (100 cases, P-value
Cl inicopathologic characteristics
MI absent (64 cases)
MI present (36 cases)
Total (100 cases)
Non-mucinous Lymph node metastases Absent Present
Number of relatives with colorectal cancers
Number of relatives without colorectal cancers
With respect to pedigree analysis, the average family
size in both RER-positive and -negative cases was 14
members (range nine to 16). A single HNPCC family
which met the Amsterdam criteria qualified for
placement in the high-grade MI group, and two HNPCC
families belonged to the Mi-negative group; MI thus
showed a 33.3% rate of sensitivity for detecting HNPCC.
In the only HNPCC family showing RER+ phenotype
cancers no hMSH2 or hMLHl mutations were identified
by the use of SSCP analysis (Dr. G. Guanti, University
of Bari, Italy, personal communication). Suspected (one
Amsterdam criteria missing) HNPCC families were
observed in all three groups: 13 in the MI negative-, one
in the low- and six in the high-MI groups, respectively.
In the group of cases with low and high MI the pedigree
analysis demonstrated a significantly (P - 0.012) higher
number (60 of 238 and 31 of 86, respectively) of relatives
affected by malignant tumors, mainly consisting of
colorectal cancers, in comparison to the group of
MInegative cases (145 of 754).
Owing to the strict similarities in clinico-pathologic
features of carcinomas with low and high MI, all RER+
cases were considered as a single group whose
pathologic and clinical parameters were then compared with
those of the group of RER-negative cases (Table 3). The
RER-positive phenotype proved to be positively
correlated with mucinous histological type (P = 0.00007),
peritumoral lymphocytic infiltration (P = 0.008),
absence of lymph node metastasis, low mitotic index (P
0.01) and a high number of relatives affected by
colorectal cancer (P = 0.03).
In this study the MI analysis, performed with 10
polymerase chain reaction primer sets on a consecutive series
of colorectal cancers, was compared with family history,
and based on genetic counselling of the patients and
with other pathological and biological characteristics.
Genomic instability was detected in 36% of the series
of 100 consecutive primary colorectal cancers examined.
We observed MI at multiple loci (more than 2 loci
altered) in 9% and MI at 1-2 loci in 27% of investigated
cancers. Genomic instability in colorectal cancers has
been reported in percentages ranging from 15% to 28%
[3, 18, 22, 23] and in all studies the frequency of MI was
higher in familial than in sporadic cases. The high
percentage of MI instability detected in our study over
those in other studies [3,18, 22 and 23] may be explained
by the high number and the location of loci investigated
to determine MI, and this technical approach may
increase the percentage of Mi-positive cases.
We found a different involvement of loci between the
group with high (MI at more than 2 loci) and the one
with low (MI at 1 or 2 loci) instability. In the first group
microsatellites mapping (D2S123, D2S119, D3S1611,
D7S481 and D7S517) near the HNPCC genes and near
or within (D13S175, APC, TP53 and D17S787) the genes
involved in cell-cycle or cell adhesion control were more
often involved. By contrast, in the low-instability group,
the more frequently unstable loci were D3S1611 and
DCC, suggesting the presence of a different type of
instability for this group of cases. For this reason, the
instability degree in this group of cancers could be
underestimated. However, the locus or allelic
heterogeneity may cause a variable degree of instability without
detectable clinico-pathological differences.
In our study the RER phenotype was correlated with
different clinico-pathological characteristics, including
age and sex of the patients, and location, size,
histological type, grade, stage, lymph node metastasis,
peritumoral lymphocytic infiltration, tumor necrosis, stromal
fibrosis, vascular invasion, neural invasion and mitotic
index of the tumors. No significant differences between
RER-positive and RER-negative cases were observed
for the majority of variables considered. A highly
significant correlation, however, was found between the
mucinous type and RER-positive phenotype of
colorectal cancers, indicating that genomic instability is
involved in the tumorigenesis of mucinous
adenocarcinomas. In addition to the mucinous aspect, the RER+
cases in our study were characterized by intense
lymphoid peritumoral infiltrates, a relatively low mitotic
index and a low incidence of lymph node metastasis. Our
findings are in agreement with data previously reported
by other authors [24, 25] and confirm that tumors with
replication errors comprise a distinct biological entity
within the colorectal cancers. In particular, the
mucinous aspect, the low proliferation rate, as determined by
analysis of S phase fraction and AgNOR content 
and the marked peritumoral inflammatory infiltrates,
probably representing host response , indicate that
RER-positive tumors are less aggressive than
RER-negative tumors. This is also confirmed by studies
suggesting that patients with RER-positive sporadic cancers
have a better prognosis than those with RER-negative
cancers [22, 23, 26].
With respect to family history, the RER+ phenotype
is significantly associated with familial clustering of
colorectal cancers but alone is not always suggestive of
the typical HNPCC. In fact, we identified one HNPCC
family in the RER+ group, and two HNPCC families in
the RER- group. Also the 'suspected' HNPCCs (one
Amsterdam criterion missing) were more frequent in
RER- than in RER+ cancer groups.
Also noteworthy was the fact that the mean ages of
patients with RER+ and RER- cancers were not
significantly different. These results are unexpected and
clearly indicate that the HNPCC cancers are not caused
by MMR-defective genes only, but that other genes may
also play a role. On the other hand, these results
correlate well with other reports from the literature indicating
the presence of HNPCC families meeting Amsterdam
criteria and lacking MMR gene mutations [28, 29].
In conclusion, our data demonstrate that RER
screening is probably not an efficient strategy for identifying
HNPCC cases, despite various relevant considerations
. RER detection is less efficient and more expensive
than pedigree analysis. Although MI in human colon
cancers is not a useful clinical indicator of HNPCC, it
appears to indicate a category of colorectal cancers with
favourable prognostic features and should be
investigated at least in all cases of mucinous colorectal
This research was supported by grants from the Italian
MURST and the AIRC (progetto speciale tumori del
1. Knudson AJ Jr. Heredity and cancer in man . Progr Med Genet 1973 ; 9 : 113 - 58 .
2. Eshleman JR , Markowitz SD . Mismatch repair defects in human carcinogenesis . Hum Mol Genet 1996 ; 5 : 1489 - 94 .
3. Aaltonen L , Peltomaki P , Leach Fet al. Clues to the pathogenesis of familial colorectal cancer . Science 1993 ; 260 : 812 - 6 .
4. Ionov Y , Peinado MA , Malkbosyan S et al. Ubiquitous somatic mutations in simple repeated sequences reveal a new mechanism for colonic carcinogenesis . Nature 1993 ; 363 : 558 - 61 .
5. Lazar V , Grandijouan S , Bognel C. Accumulation of multiple mutations in tumor suppressor genes during colorectal tumorigenesis in HNPCC patients . Hum Mol Genet 1994 ; 3 : 2257 - 60 .
6. Markowitz S , Wang J , MyerofF L et al. Inactivation of the type II TGF-p receptor in colon cancer cells with microsatellite instability . Science 1995 ; 268 : 1336 - 8 .
7. Pearsons R , Myeroff LL , Liu B et al. Microsatellite instability and mutations of the transforming growth factor p type II receptor gene in colorectal cancer . Cancer Res 1995 ; 55 : 5548 - 50 .
8. Bronner CE , Baker SM , Morrison PTet al. Mutation in the DNA mismatch repair gene homologue hMLHl is associated with hereditary non-polyposis colon cancer . Nature 1994 ; 36 : 258 - 61 .
9. Nicolaides N , Papadopoulos N , Liu B et al. Mutation of two PMS homologues in hereditary nonpolyposis colon cancer . Nature 1994 ; 371 : 75 - 80 .
10. Leach FS , Nicolaides NC , Papadopoulos N et al. Mutation of a MutS homolog in hereditary nonpolyposis colorectal cancer . Cell 1993 ; 75 : 1215 - 25 .
11. Papadopoulus N , Nicolaides NC , Wei YF et al. Mutation of a mutL homolog in hereditary colon cancer . Science 1994 ; 263 : 1625 - 9 .
12. Miyaki M , Konishi M , Muraoka M , Kikuchi-Yanoshita R et al. Germ-line mutation of hMSH2 andhMLHl genes in Japanese families with hereditary non-polyposis colorectal cancer (HNPCC): Usefulness of DNA analysis for screening and diagnosis of HNPCC patients . J Mol Med 1995 ; 73 : 515 - 20 .
13. Peltomaki P , Lothe RA , Aaltonen LA et al. Microsatellite instability is associated with tumors that characterize the hereditary non-polyposis colorectal carcinoma syndrome . Cancer Res 1993 ; 53 : 5853 - 5 .
14. Brown ML , Kessler LG . The use of gene tests to detect hereditary predisposition to cancer: Economic considerations . J Natl Cancer Inst 1995 ; 87 : 1131 - 6 .
15. Brown ML , Kessler LG . Use of gene tests to detect hereditary predisposition to cancer: What do we know about cost effectiveness? Int J Cancer 1996 ; 69 : 55 - 7 .
16. Vasen HFA , Mecklin JP , Khan PM et al. The International collaborative group on hereditary non-polyposis colorectal cancer (ICG-HNPCC) . Dis Colon Rectum 1991 ; 34 : 424 - 5 .
17. Jass JR , Cottier DS , Jeevaratnam P et al. Diagnostic use of microsatellite instability in hereditary non-polyposis colorectal cancer . Lancet 1995 ; 346 : 1200 - 1 .
18. Samowitz WS , Slattery ML , Kerber RA et al. Microsatellite instability in human colonic cancer is not a useful clinical indicator of familial colorectal cancer . Gastroenterology 1995 ; 109 : 1765 - 71 .
19. Jass JR , Sobin LH , Histological Typing of intestinal Tumors . Second edition. Berlin: Springer Verlag 1989, 32 - 3 .
20. Hermanek P , Sobin LH . TNM Classification of malignant tumors . Fourth edition , second revision. Berlin: Springer-Verlag 1992 .
21. Benatti P , Sassatelli R , Roncucci L et al. Tumor spectrum in hereditary non-polyposis colorectal cancer (HNPCC) and in families with 'suspected H N P C C A population-based study in Northern Italy . Int J Cancer 1993 ; 54 : 371 - 7 .
22. Lothe RA , Peltomaki P , Meling Gi et al. Genomic instability in colorectal cancer: relationship to clinicopathological variables and family history . Cancer Res 1993 ; 53 : 5849 - 52 .
23. Thibodeau SN , Bren G , Schaid D et al. Microsatellite instability in cancer of the proximal colon . Science 1993 ; 260 : 816 - 9 .
24. Bocker T , Schlegel J , Kullmann F et al. Genomic instability in colorectal carcinomas: Comparison of different evaluation methods and their biological significance . J Pathol 1996 ; 179 : 15 - 9 .
25. Kim H , Jen J , Vogelstein B et al. Clinical and pathological characteristics of sporadic colorectal carcinomas with DNA replication errors in microsatellite sequences . Am J Pathol 1994 ; 145 : 148 - 56 .
26. Lynch HT , Smyrk TC , Watson P et al. Genetics , natural history, tumor spectrum, and pathology of hereditary nonpolyposis colorectal cancer: An update review . Gastroenterology 1993 ; 104 : 1535 - 49 .
27. Moslein G , Tester DJ , Lindor NM et al. Microsatellite instability and mutation analysis of hMSH2 and hMLHl in patients with sporadic, familial and hereditary colorectal cancer . Hum Mol Genet 1996 ; 5 : 1245 - 52 .
28. Dunlop MG , Farrington SM , Carothers AD et al. Cancer risk associated with germline DNA mismatch repair gene mutations . Hum Mol Genet 1997 ; 6. 105 - 10 .
Received 5 March 1998; accepted 27 May 1998 .