Conference 2008 Roskilde

Friday Report

Clinical Management

See also Abstracts -- Thursday report -- Saturday report

Eamonn Maher of the UK reported on aspects of tumor development in VHL. He began by describing the classification determined in 1994 of VHL Types 1, 2A, 2B, and 2C.  We now know that it’s more complicated than originally thought.  In fact the risks of the various tumors do vary, but none of them can with confidence can be thought to be zero.  In the 2C classification, for example, which was thought to be “pheo only,” some family members may still develop the other tumors.  It is better to say that the incidence of these other aspects is low, but not zero.  It is still important to screen for all aspects of VHL, though one might vary the protocol a bit to reflect the lower probability of some aspects.

One of the advantages of the classification scheme is that his group as checked them in vitro to see the variations in the ability of the different mutation types to regulate HIF.  From this they were able to determine that the risk of pheo is linked to the ability to regulate HIF.  In 2C the pathway did not involve ability to regulate HIF.  Rather the VHL protein also plays a role in apoptosis, which provides a HIF-independent pathway to pheochromocytoma.

The risk of hemangioblastoma seems related to the ability of VHL to dysregulate HIF.

Age-related risk of retinal angiomas seems to be higher with truncating mutations then for germline deletions and Loss of Function (LOF) missense substitutions.

His group also checked people in the UK who had been diagnosed as having sporadic hemangioblastomas of the Central Nervous System (CNS) (198 patients) and determined that 4% of them in fact had germline mutations of VHL.  A number of people who did not have a germline VHL mutation eventually went on to have another VHL-associated tumor, indicating that they might be mosaic for VHL.

For those who did have an underlying VHL mutation, two patients had Chuvash polycythemia (who were not from Chuvashia), and one had a C162Y mutation, previously reported in Japan, with CNS hemangioblastomas and no other aspects.

People with VHL Type 1 and Type 2B mutations have a higher incidence of renal cell carcinoma (RCC).  Those with truncating mutations had an earlier mean age of symptoms (38.2 years) than those with deletions (45.8 years) or LOF missense mutations (42.3 years).  Dr. Maher hypothesizes that possibly the 3’ deletion involving HSPC300 has a protective effect against RCC? Or that there are “dominant negative effects” of truncated pVHL.

He also looked through the UK registry to find people with familial pheochromocytoma, where they found that 40% of these patients proved to have an underlying VHL mutation. 

Among people in the registry who had familial clear cell renal cancer (ccRCC), he found underlying VHL, 4% BHD (with a folliculin mutation), and a few SDHB mutations.

There is an overlap between SDH and VHL:

• SDH mutations cause an accumulation of succinate that inhibits PD enzymes and so upregulate HIF
• So HIF overexpression is associated with both VHL and SDH inactivation
• It also interrupts the pathway to apoptosis, causing the preservation of pheo precursor cells

He therefore looked for similarities in the two groups of tumors, following the HIF-1, HIF-2, and the HIF-independent pathways.

He found differences in gene expression patterns, and wondered whether SDH mutations might cause an RCC-only phenotype.  He found one family with familial RCC (ages 24, 24, and 73) and two isolated cases of bilateral RCC at ages 30 and 38 years.  There was no evidence of pheo or head and neck paragangliomas among these patients or their relatives.  All three of these mutations were in the first 50 amino acids.

He thus calculated in SDHB a lifetime risk of HNPGL 80%, and of RCC 11%

There is evidence for genetic modifiers in VHL.  Ten years ago e did a study comparing the frequency of retinal angiomas, and found more similarity among close relatives than would have been expected.  He identified MMP3 as a modifier of RCC, and MMP1 might influence the risk of hemangioblastomas of eye and brain.  Genotyping of functional polymorpisms in VHL target genes revealed that genetic variants in MMP3 and MMP1 influence phenotypic expression of VHL disease.

Phenotypic variation is also due to environmental factors.  Cancer is a multi-step process, so far best described in the case of colorectal caner.  For VHL we do not yet have a good idea of the chain of events that occur after the loss of the second copy of VHL. 

Jose Claudio Rocha of Brazil reported his findings of VHL germline mutations in Brazil.  He has been studying VHL in Brazil since 1998, and as located 43 families.  Families are of mixed Portuguese, African, and Brazilian native ethnicity.  50% of these families have de novo mutations.  29 are point mutations, while 5 are large deletions.  Point mutations seem to concentrate in exon 1, and he sees very little pancreatic involvement.

Mette Dandanell Nielsen of Denmark reported on the germline mutations found in Denmark.  94 individuals from 29 families were tested in Copenhagen, along with 166 individuals with no family history.  They found no evidence of a founder mutation in Denmark.

Ashok Pillai of India reported his findings of VHL in Kerala State in southwestern India.  He works in a 10-year-old medical referral center in Kerala.  In the literature in India there are so far fewer than five case reports.  He is working to establish high quality DNA testing for VHL in India.

So far he has identified four families in Kerala, two of whom were lost to follow-up due to the stress of screening.  Of these 69% have hemangioblastomas, 31% RCC, 100% have multicystic pancreas, and none of them has any retinal angioma, no polycythemia, and no pheos.

In the literature he found that among cases of complete deletions of pVHL there was only 7% retinal involvement.  There is a possibility of a founder effect among these four families, with a large deletion explaining the lack of retinal involvement.

His team is offering stereotactic radiosurgery for hemangioblastomas and nephron-sparing surgery for RCC.  He is hoping to spread more information about VHL in Kerala, where there are 30 million people.  There is nearly 100% literacy in this state, so distribution of this kind of information is easier in this state than in many others. 

Lesley Andrews of Australia reported that the first cancer genetics practice in Australia was established in 1994.  Now they see about 900 cases each year, but few VHL families.  They are working to reduce the time-to-diagnosis of VHL in Australia.  At this point, the average age at first symptom is 20.6 years and the average age at diagnosis is 25.3 years. 

They are routinely screening people who meet the diagnostic criteria for VHL.  They have changed from 24-hour urine to plasma metanephrines, and find that with this more efficient test people are more compliant with this screening requirement.  They have also moved from separate imaging of CNS and abdomen to a single MRI.

Interested specialists at the Prince of Wales Hospital (POWH) are happy to develop expertise in VHL, but logistically they have not been able to see patients at a single clinic appointment.  They are hoping to evolve their activities toward a centralized care model with a state-wide registry, information, support, a screening reminder service, and research resource.

The clinical genetics service at POWH is developing a questionnaire on psychsocial issues, and hopes to apply it internationally.  Lesley will speak about this on Saturday.

Hartmut Neumann of Germany spoke on familial pheochromocytoma.  He has collected germline mutations in 1331 patients with symptomatic pheochromocytoma and/or paraganglioma, and 271 non-syndromic pheos.

RET 7%
VHL 13%
SDH 6%
SDH 2%
SDHD 6%
NF1 4%
Total with any mutation 37%

Large deletions represent 1% of these 1331 cases.  Large deletions are confirmed with two different methods.

He strongly recommends laparoscopic partial adrenalectomy.  Martin Walse in Essen is the best endoscopic surgeon in Germany.  It would be best to have 5-6 such centers in Germany.

Of people with Glomustumor (German for head and neck paraganglioma or HNPGL), the majority are SDHB mutations.  Of all pheos, the majority are VHL mutations.  Of those with multiple pheos, or extra-adrenal pheos, the majority are VHL (41%), SDHB=15%

Of thoracic pheos, 23% VHL; 21% SDHB; 21% SDHD
Of malignant pheos, 16% VHL; 25% SDHB; 9% NF1; 50% sporadic

In determining who might have hereditary pheo, age is the best indicator.  90% of patients under age 10 will have a hereditary risk factor; 20% of those age 60 will have a mutation.  There is a linear progress between those two points.  For the first four decades, the probability of VHL is the highest.

He spoke about VHL Type 2C.  In short, he does not like this category “pheo only.”  Rather, he agrees with Dr. Maher that one should say “predominantly pheo” since occasionally other tumors do occur. 

In his series in Germany, out of 174 patients, they presented with only pheo and no additional lesions.  However on follow up only 97 of these still had no additional lesions; 77 patients had developed additional VHL lesions over time.  36 of these people had no prior family history.  In order to gain some perspective on these people, if we add other patients with these mutations from other papers, the number of other lesions grows:
Mutation c.550 = 29/53 have only pheo
c.680, 7/11 have only pheo
c.695, 6/11
c.712  9/34
c.713, 1/18
c.775, 1/13

He feels that c.505 mutations should be reclassified as Type 2B.

HNPGL in VHL are rare (n=3 out of a series of 382 patients in European network for HNPGL).  Neumann tests all people with pheos for all five genes that can cause pheo.

Malignant pheos often have distant metastases.  They require scintigraphic imaging and screening for elevated catecholamines.  SDHD patients are younger, otherwise there are no clear differences.

Giuseppe Opocher of Italy reported to us about paraganglioma in Trentino

Ten unrelated people in the area around Trentino, Italy, were found to have the same SDHD mutation for HNPGL.  Dr. Opocher checked the birth and death recods and was unable to find any relationship going back 200 years.  Nonetheless it was fairly evident that there had to be a founder effect. 

Through a series of calculations and historical investigations he was able to determine that it was likely a family of German origin that immigrated to the region about 1400 when some mines were discovered.  This same mutation has also been seen in Germany and Poland, which may be other branches of this same family.

Julia Forman of England reported her analysis of 45 mutations among 261 patients in England.  She found few predictive factors, the most notable of which are

• 15 mutations in domain 1, which binds to HIF, leading to RCC
• 8 mutations in domain 2, which binds to elongins B and C

All mutations with HIF interactions cause Retinal angiomas (RA), Capillary hemangioblastomas (CHB), and renal cell carcinoma (RCC).  She then did further analysis to determine which mutations most destabilize that first domain.  She suspects that disruption of the P53 binding is also a factor.

She concludes that there are two separate mechanisms whereby VHL mutations cause Renal cell carcinoma:

• Disrupting the HIF binding, directly or through destabilization of the binding domain, causing RA/CB/RCC
• Mutations at the Elongin B binding site.

Nicola Bradshaw of Scotland reported her findings among a large Scottish kindred with variable penetrance.  This mutation has been published several times, and does not always carry a low rate of penetrance.  She hypothesizes that in this family an additional genetic alteration, Ser65Leu, disrupts the coordination of connection between VHL and HIF-1alpha and may compensate for mutation in the L1 domain of VHL

Sophie Giraud of France reported that she has developed a method for verifying somatic mosaicism.  She worked with two patients who had clinical manifestations of VHL, but whose mutations could not be found in the blood. 

She analyzed these two patients by dHPLC.  In each case she located some aberrant peaks, which she they analyzed further using the dHPLC collector.  She then collected and sequenced separately the various peaks, clearly identifying two nucleotides of equivalent heights, allowing her to confirm the diagnosis of VHL somatic mosaicism. 

For both these patients, VHL follow-up will be done, and DNA analysis of their offspring will be offered.

Carsten Brandt of Denmark presented data from Denmark and asks whether cases of hemangioblastoma of the brain and including hemangioblastomas of the retina might be a separate entity, which he suggests might be called “hereditary hemangioblastomas”.

There was little agreement with this proposal.  Dr. Maher suggested a possible misdiagnosis of Coats’ disease.   Dr. Wiklund suggested that hemangioblastoma in the pons might be a cavernous hemangioma rather than a hemangioblastoma of VHL.  Dr. Rocha questioned whether retinal hemangioblastoma and visceral cysts should be considered diagnostic of VHL. 

In Summary, Dr. Maher thanked all the speakers for a fascinating morning.  We have learned a great deal, but clearly there is still much more to learn.  Hopefully we may be able to explain more about these differences in the coming years.

See also Abstracts -- Thursday report -- Saturday report