Clinical and molecular characterization of BRCA-associated breast cancer: results from the DBCG

Abstract

Background: In breast cancer (BC) patients a cancer predisposing BRCA1/2 mutation is associated with adverse tumor characteristics, risk assessment and treatment allocation. We aimed to estimate overall- (OS) and disease-free survival (DFS) according to tumor characteristics and treatment among women who within two years of definitive surgery for primary BC were shown to carry a mutation in BRCA1/2 .

Material and methods: From the clinical database of the Danish Breast Cancer Group we included 141 BRCA1 and 96 BRCA2 BC patients. Estrogen receptor and HER2 status were centrally reviewed on paraffin-embedded tumor tissue. Information on risk reducing surgery was obtained from the Danish Pathology and Patient Registries and included as time-dependent variables in Cox proportional hazard models.

Results: Ten-year OS and DFS for BRCA1 BC patients were 78% (95% CI 69–85) and 74% (95% CI 64–81). Ten-year OS and DFS for BRCA2 BC were 88% (95% CI 78–94) and 84% (95% CI 74–91). BRCA1 BC patients as compared to BRCA2 BC patients had a higher risk of BC relapse or non-breast cancer within ten years of follow-up, independent of ER status (adjusted HR 2.78 95% CI 1.28–6.05, p = .01), but BRCA mutation was not associated with OS (adjusted HR 1.98, 95% CI 0.87–4.52, p = .10). In multivariate analysis, including both BRCA1 and BRCA2 carriers, no chemotherapy was associated with a higher risk of death (adjusted OS HR 3.58, 95% CI 1.29–9.97, p = .01) and risk reducing contralateral mastectomy (RRCM) was associated with a significantly reduced risk of death (adjusted OS HR 0.42, 95% CI =0.21–0.84, p = .01).

Conclusion: Difference in OS between BRCA1 and BRCA2 BC patients could be ascribed to tumor-biology. BRCA1 BC patients may have a shorter ten-year DFS than BRCA2 BC patients. Chemotherapy and risk reducing contralateral mastectomy reduce mortality for both BRCA1 and BRCA2 BC patients.

Background

In around 80% of BRCA carriers, breast cancers (BC) will be high grade invasive ductal carcinomas of no special type (NST) [1–3]. Medullary features and abundance of lymphocytes in the tumor-microenvironment is more often present in BRCA1 carriers and less than 25% express estrogen (ER) or progesterone receptors (PR) and only around 10% express human epidermal growth factor receptor 2 (HER2) [4,5]. The majority of BRCA2 BCs are ER/PR positive. Compared with sporadic BC a higher proportion are high grade but without other histopathological characteristics different from sporadic BC [4,5]. Identification of a cancer predisposing BRCA mutation at the diagnosis of BC may influence decisions about treatment and cancer prophylaxis. Recent moderate-sized population-based studies have shown a higher mortality in BRCA BC patients if chemotherapy is omitted, indicating that allocation of chemotherapy by standard guidelines have not been optimal for this subgroup of BC patients [6,7]. However, risk reducing salpingo-oophorectomy (RRSO) and risk reducing contralateral mastectomy (RRCM) are recommended in BRCA1/2 mutation carriers since they are associated with reduced risk of BC and ovarian cancer and a better overall survival (OS) [8].

Here we report ten-year survival estimates and associates of treatment-, patient- and tumor characteristics with survival in a large cohort of BRCA1/2 BC patients with detailed prospectively collected clinical data from the national database of The Danish Breast Cancer Group (DBCG).

Method and materials

Patients

The Danish Data Protection Agency (2009-41-3611) and the Danish Ethical Committee (registration number 33483) approved the study.

Danish patients are registered for health care and administrative purposes by a unique personal identification number (CPR) in national clinical medical registries [9]. The CPR was used to cross-link information from registries. Hospital admissions, outpatients, surgery and diagnosis have been registered in The Danish National Patients Registry (DNPR) [10] since 1977, pathological procedures in the Danish Pathology Register (DPR) since 1997 [11] and cancer incidences in The Danish Cancer Registry (CR) since 1943 [12]. Since 1977, Danish women with primary BC have been registered in the DBCG database including prospectively registered clinical data. Family history of breast- and ovarian cancer was added in the Hereditary Breast and Ovarian Cancer (HBOC) register in 1999 [13]. Patients suspected to carry a high-risk mutation and referred for genetic counseling were registered with a unique family-number. National guidelines from 2001 proposed referral for genetic counseling in families with two first-degree relatives with BC under the age of fifty or ovarian cancer, three first degree relatives with BC, breast and ovarian cancer in the same individual and male BC. For high-risk patients undergoing mutational screening, the finding of a deleterious BRCA1/2 mutation was subsequently registered [14,15].

By standard clinical guidelines [16] early stage I–III BC patients undergo mastectomy or breast-conserving surgery and SN procedure if clinical node-negative, followed by radiation therapy and adjuvant systemic treatment. Patients older than 60 at diagnosis with small tumor size (≤10 mm), node negative, ER positive (≥10%), HER2 negative and grade I ductal or grade I–II lobular tumors are classified as low risk and not eligible for systemic treatment. This group amounts to approximately 8% and consequently approximately 90% of early BC patients receives systemic treatment. Trastuzumab was introduced in 2005 [17] and taxanes in 2007 [18]. Over time, surgery has become less extensive with breast-conserving surgery, smaller margins and introduction of the SN procedure in 2001 [17,19]. However, primary mastectomy was recommended since 2001 in BC patients with BRCA1/2 mutation [14].

Patients had clinical follow-up every three months, the first year after BC, every sixth months, two to five years after BC and once a year, 6–10 years after BC or until a first event. Death before clinical follow-up, distant- or local relapse, contralateral BC or non-breast cancer were registered as an event.

BRCA1/2 mutational status was updated until July 2011 and BC surgery until February 2012. Data on clinical follow-up, vital status, treatment, surgery and pathology were updated until September 2016. Information about vital status by linkage to the Danish Civil Registration System [9] is complete until September 2016.

Patients were excluded if they had BRCA1/2 mutation screening performed more than two years after date of definitive surgery for primary BC, if they had stage four disease at diagnosis, previous malignant disease (other than non-melanoma skin cancer) or no clinical follow-up.

RRSO was defined as removal of both ovaries. RRCM was defined as bilateral mastectomy before BC diagnosis, bilateral mastectomy after primary breast-conserving surgery or contralateral mastectomy after primary ipsilateral mastectomy. Dates of operation for RRSO and RRCM were retrieved from DNPR by searching ICD8/ICD10 codes for prophylactic procedures combined with codes for surgical removal of ovaries or mastectomy. Search in DPR was performed during April and May 2017 to review pathology reports from all patients not found to have risk reducing surgery in DNPR, if there was uncertainty of indication for mastectomy/salpingo-oophorectomy or uncertainty about ‘site’ of RRSO/RRCM. Mastectomy or salpingo-oophorectomy associated with a malignant diagnosis were not considered. If no information was found in the registers, we assumed the patient did not have risk reducing surgery. Information on other malignancy during follow-up was registered in the DBCG database, however date of primary ovarian-, fallopian tube- or peritoneal cancer during and after end of clinical follow-up was retrieved from CR. Data from CR were updated until 31 December 2014.

Among 1507 BRCA1/2 mutation carriers, 597 were diagnosed with primary invasive BC and 583 had definitive surgery. Median time from BC to BRCA testing was 0.7 years (240 days and 251 days for BRCA1 and BRCA2, respectively). Forty-three patients were tested before BC diagnosis (twenty-seven BRCA1 and sixteen BRCA2). In 285 patients the BRCA1/2 mutational screening was performed less than two years after surgery. Three patients with stage IV disease, eighteen with previous malignancy and twenty-seven patients with no clinical follow-up were excluded, resulting in 237 BC patients with a BRCA1/2 mutation (141 BRCA1 and 96 BRCA2) included for survival analysis (Figure 1).

Figure 1: Patients flowchart.

BRCA1/2 screening

BRCA1/2 mutational screening was offered as a diagnostic test from 1997 in East Denmark and from 1999 in Western Denmark. BRCA1/2 mutation screening and mutation variant classification was performed in three centers as previously described [20]. Methods used were denaturing high performance liquid chromatography or temperature gradient electrophoresis and protein truncation test. Validation was performed with Sanger sequencing, or Sanger sequencing and multiplex ligation-dependent probe amplification.

IHC/ISH analysis and tumor morphology

Immunohistochemical (IHC) and in situ hybridization analysis was performed on formalin-fixed paraffin-embedded (FFPE) tumor tissue to obtain accurate ER and HER2 status. FFPE blocks with primary BC tissue were collected from Danish pathology departments. IHC and ISH analysis were performed on tissue-micro-arrays (TMA) with up to four 1.5 mm cores from each tumor. Four micrometer sections from TMA blocks were stained with validated antibodies for ER (clone SP1 RTU with visualization system Ultraview or Optiview, Ventana) and HER2 (clone 4B5 RTU, visualization system Ultraview, Ventana). The tumor was reported as ER positive if ≥1% of positive nuclear staining. HER2 IHC was scored according to standardized guidelines [21]. In HER2 equivocal cases, HER2 gene amplification was determined with a gene-protein assay (INFORM Dual ISH DNA Probe cocktail, Ventana) and scored according to guidelines [21]. Malignancy grade and histological subtype were determined with standard light microscopy on full-size four micrometer sections from HE stained FFPE tumor sections. Grading was performed according to the modified Bloom-Richardson-Elston system [22] and histological subtype according to WHO classification of tumors [23].

Statistical analysis

Disease-free survival (DFS) was defined as the time from definitive BC surgery to invasive loco-regional recurrence, distant metastasis, contralateral invasive breast cancer, second primary non-breast invasive cancer or death irrespective of cause. Overall survival was calculated as elapsed time from BC surgery to death. DFS and OS were analyzed unadjusted by the Kaplan-Meier method and the log-rank test to compare groups. Considering contralateral BC, local- and distant relapse and death as first event competing events, ten-year cumulative incidence of non-breast cancer was calculated in competing risk analysis.

Hazard ratios (HR) were estimated from the Cox proportional hazard regression model to quantify effects and to explore interactions. The assumption of proportional hazards (PH) was tested by plotting Schoenfeld residuals against time and by testing correlation between ranked failure times and Schoenfeld residuals. A time-dependent component was included for ER status to comply with proportional hazards. BRCA status (BRCA1 vs. BRCA2), ER status (positive vs. negative), positive lymph node metastases (no vs. yes), adjuvant chemotherapy (yes vs. no) and RRCM were included in multivariate analysis. RRSO and RRCM were included as time-dependent covariates. Missing values were imputed with the most frequent value in the remaining patients. Sensitivity analysis was performed with a model where patients with missing values were excluded and the resulting hazard ratios were not significantly different. We used two-sided p value with 5% significance level.

Differences of distribution of patient and tumor characteristics between subgroups were assessed with Chi-square test or Fishers exact test where appropriate.

Results

Treatment, patient- and tumor-characteristics are presented in Table 1. ER, PR and HER2 status were obtained from the central pathology review or retrieved from DBCG clinical database (ER in ten and HER2 in eight patients). Mean age at diagnosis was 42 and 46 years for BRCA1 and BRCA2 carriers, respectively. The majority were premenopausal at diagnosis with invasive ductal and HER2 negative tumors. BRCA1 tumors were ER negative/TN in contrary to BRCA2 tumors which were mainly ER positive. The majority of BRCA1 tumors were high grade and lymph node negative while almost two thirds of BRCA2 patients were lymph node positive. A mastectomy was the primary surgery in just over half irrespective of the involved gene but the uptake on RRCM was somewhat higher in BRCA1 carriers (Table 1). The vast majority received adjuvant chemotherapy and RRSO. Median time to RRSO and RRCM after BC surgery was 1.3 and 1.6 years (minimum 12 and zero days respectively to maximum 10.6 years) for both BRCA1 and BRCA2 patients, respectively.

Table 1. Patient and tumor characteristics.

	Patients
	BRCA1		BRCA2
	n	%	n	%
Variable	141	59	96	41
Age
18–39	66	47	31	32
40–60	62	44	52	54
>60	13	9	13	14
Year of diagnosis
1998–1999	1	1	1	1
2000–2004	29	21	20	21
2005–2009	86	61	58	60
2010–2011	25	18	17	18
Menopausal status at diagnosis
Premenopausal	111	79	64	67
Postmenopausal	29	21	32	33
Unknown	1	1	0	0
Histology
Ductal	129	91	85	89
Lobular	0	0	5	5
Other	12	9	6	6
Tumor size
≤10 mm	19	13	17	18
11–20 mm	64	45	40	42
21–50 mm	56	40	36	38
>50 mm	2	1	3	3
Grade
1	3	2	8	8
2	29	21	40	42
3	97	69	41	43
Non-ductal/non-lobular histology	12	9	6	6
Missing	0	0	1	1
Positive lymph nodes
0	82	58	34	35
1–3 positive	41	29	47	49
≥ 4 positive	18	13	14	15
Unknown	0	0	1	1
Bilateral breast cancer
Yes	2	1	3	3
No	136	97	92	96
Unknown	3	2	1	1
ER status
Positive	44	31	76	79
Negative	97	69	20	21
PR status
Positive	35	25	67	70
Negative	106	75	29	30
HER2 status
Positive	6	4	6	6
Negative	134	95	89	93
Unknown	1	1	1	1
TN
Yes	91	65	19	20
No	50	35	77	80
Surgery
Mastectomy	76	54	54	56
Breast conserving surgery	65	46	42	44
Systemic therapy
None	2	1	1	1
Chemotherapy alone	106	75	23	24
Chemotherapy and endocrine therapy	28	20	59	61
Endocrine therapy alone	5	4	13	14
Radiation therapy
Yes	84	60	72	75
No	42	30	16	17
Unknown	15	11	8	8
RRCM
Yes	95	67	52	54
No	46	33	44	46
RRSO
Yes	132	94	90	94
No	9	6	6	6
RRCM only
Before BC surgery	0	0	0	0
≤2 years after BC surgery	3	2	1	1
>2 years after BC surgery	2	1	0	0
RRSO only
Before BC surgery	3	2	6	6
≤2 years after BC surgery	26	18	28	29
>2 years after BC surgery	13	9	5	5
RRSO and RRCM
Before BC surgery	1	1	0	0
≤2 years after BC surgery	66	47	33	34
>2 years after BC surgery	23	16	18	19

ER: estrogen receptor; HER2: human epidermal growth factor Receptor 2; PR: progesterone receptor; TN: triple negative, RRCM: risk reducing contralateral mastectomy; RRSO: risk reducing salpingo-oophorectomy.

Median potential follow-up time for OS and DFS was 9.01 (BRCA1 9.01, BRCA2 8.52) and 7.15 years (BRCA1 7.02, BRCA2 7.15), respectively. Of BRCA1 patients, eighteen (13%) had local- or distant recurrence, five (4%) had contralateral breast cancer, three died before follow-up (2%) and seven had non-breast cancer (5%). Nine (9%) BRCA2 patients had local- or distant relapse, one died before clinical follow-up (1%) and two had non-breast cancer (2%). Ten-year cumulative incidence of local- or distant relapse, contralateral BC and death as first event or non-breast cancer was 19.0 (95%CI 13–27) and 6.9% (95% CI 2.9–13.4) for BRCA1 and 13.4 (95% CI 6.6–22.6) and 2.2% (95% CI 0.4–7.0) for BRCA2 carriers, respectively. A higher proportion of events in BRCA1 carriers were non-breast cancer (36%) compared to BRCA2 carriers (16%). Two patients had ovarian cancer, both BRCA1 carriers. Cancers other than BC and ovarian cancer were lung cancer (one BRCA2 carrier), gastrointestinal cancer (two BRCA1 carriers, one BRCA2 carrier), malignant melanoma (two BRCA1 carriers) and unknown type (one BRCA1 carrier). DFS was significantly worse in BRCA1 compared to BRCA2 carriers (Figure 2, Table 3) and adjusted for lymph node status, adjuvant chemotherapy and RRCM, BRCA1 mutation was associated with a significantly higher hazard (HR 2.90 95% CI 1.4–5.9, p = .003). When ER was included in the model the association was still significant (HR 2.78 95% CI 1.28–6.05, p = .01). BRCA1/2 status was not statistical significantly associated with higher risk of death, however a trend towards worse OS was observed.

Figure 2: (a,b) DFS and OS according to BRCA1/2 status; (c,d) Disease free- and overall survival according to adjuvant treatment with or without chemotherapy.

RRCM was not statistical significantly associated with DFS, however, significantly lower risk of death was observed in both unadjusted and adjusted analysis. Eighteen percent of patients that did not have RRCM versus three percent of patients that did, were older than 60 years at BC diagnosis.

Of patients who had previously undergone RRSO, one patient was diagnosed with primary peritoneal cancer 14.6 years after. One patient with intact ovaries had ovarian cancer. There were no incidences of ovarian cancer after an event of relapse or contralateral BC. No association of RRSO with DFS or OS was observed in univariate analysis.

Patients not treated with chemotherapy showed a trend towards worse outcome for both DFS and OS (Table 2 and Figure 2) and in multivariable analysis a statistical significant effect was seen for OS, with a similar trend for DFS although not significant (p = .06; Table 3). Patients not treated with chemotherapy were >10% ER positive postmenopausal and over 50 years, except one 48-year old premenopausal patient.

Table 2. Ten-year OS and DFS according to tumor- and patient characteristics.

	BRCA1/2
	OS	DFS
	10 year, % (95% CI)	10 year, % (95% CI)
All	82 (76–87)	78 (71–83)
BRCA mutation
2	88 (78–94)	84 (74–91)
1	78 (69–85)	74 (64–81)
ER
Positive	85 (75–91)	79 (68–86)
Negative	80 (70–87)	77 (67–84)
HER2
Negative	82 (76–88)	79 (72–84)
Positive	83 (46–95)	64 (30–85)
TN
No	85 (75–91)	78 (69–85)
Yes	80 (70–87)	77 (67–85)
Positive lymph nodes
No	86 (76–92)	82 (72–89)
Yes	79 (69–86)	73 (64–81)
Tumor size
0–20 mm	87 (79–92)	80 (70–86)
>20mm	75 (63–84)	75 (65–83)
Grade
1	83 (27–97)	86 (33–98)
2	82 (66–91)	80 (68–88)
3	82 (73–88)	75 (66–82)
Not ductal/lobular	89 (64–97)	88 (61–97)
Chemotherapy
No	74 (49–89)	67(39–84)
Yes	83 (76–88)	79 (72–85)

ER: estrogen receptor; HER2: human epidermal growth factor receptor 2; TN: triple negative; OS: overall survival; DFS: disease free survival.

Table 3. Prognostic impact of tumor- and patient characteristics.

	Overall survival
	Univariate		Multivariate
	HR (95% CI)	p value	HR (95% CI)	p value
BRCA mutation
21	1 1.87 (0.91–3.87)	.09	1 1.98 (0.87–4.52)	.10
ER < seven years follow-up
Positive Negative	1 2.3 (1.05–5.15)	.03	1 2.81 (1.09–7.27)	.03
ER > seven years follow-up
Positive Negative	1 0.5 (0.13–2.06)	.33	–	–
HER2
Negative Positive	1 0.91 (0.22–3.82)	.90	–	–
Positive lymph nodes
No Yes	1 1.54 (0.79–3.00)	.20	1 1.86 (0.94–3.68)	.08
Tumor size
0–20 mm >20 mm	1 1.84 (0.96–3.52)	.07	–	–
Chemotherapy
Yes No	1 1.86 (0.77–4.47)	.20	1 3.58 (1.29–9.97)	.01
RRCM
No Yes	1 0.43 (0.22-0.85)	.01	1 0.42 (0.21–0.84)	.01
RRSO
No Yes	1 0.74 (0.29–1.89)	.54
	Disease free survival
BRCA mutation
2 1	1 2.02 (1.04–3.90)	.04	1 2.78 (1.28–6.05)	.01
ER < seven years follow-up
Positive Negative	1 1.36 (0.74–2.50)	.32	1 1.11 (0.53–2.30)	.79
ER > seven years follow-up
Positive Negative	1 0.61(0.05–6.91)	.69	–	–
HER2
Negative Positive	1 1.67 (0.60–4.66)	.36	–	–
Positive lymph nodes
No Yes	1 1.66 (0.91–3.03)	.10	1 2.14 (1.14–4.00)	.02
Tumor size
0–20 mm >20 mm	1 1.41 (0.79–2.54)	.25	–	–
Chemotherapy
Yes No	1 1.64 (0.69–3.88)	.29	1 2.48 (0.95–6.48)	.06
RRCM
No Yes	1 0.72 (0.39–1.32)	.28	1 0.70 (0.38–1.29)	.25
RRSO
No Yes	1 0.91 (0.42–2.00)	.82	–	–

ER: estrogen receptor; HER2: human epidermal growth factor Receptor 2; PR: progesterone receptor; TN: triple negative, RRCM: risk reducing contralateral mastectomy; RRSO: risk reducing salpingo-oophorectomy.

Discussion

In this nationwide study of BRCA-associated BCs in Denmark, BRCA1 carriers were shown to have a significantly shorter DFS compared to BRCA2 carriers after primary BC surgery. The majority underwent RRSO and around half RRCM after BC diagnosis. Risk reducing contralateral mastectomy significantly improved survival, but not RRSO. RRCM was associated with a more than 50 percent reduction in risk of death. The uptake of RRCM has increased through the years and our results confirm that BRCA1/2 positive BC patients benefits from RRCM after BC. However, patients not undergoing RRCM were older, which of course influences OS and age could have confounded the results to some extent. There was no association with DFS which is surprising since RRCM is associated with and mediated by lower risk of contralateral BC [8]. Finally, omission of chemotherapy was in this cohort of BRCA carriers with BC patients associated with a higher mortality.

A strength of our study is that clinical data were registered prospectively with detailed information on treatment and follow-up. Secondly we did central review of ER and HER2 status on tissue from 99% of all patients included, resulting in ER/HER2 status with standardized methods. Thirdly we retrieved information on the individual level from the DNPR and DPR registries containing complete information on risk reducing surgery.

There are limitations in the sense that even a nationwide study is restricted by sample size. The study-design contains a potential risk of bias however, we restricted the risk of longevity bias by excluding patients identified as a BRCA carrier more than two years after BC diagnosis. Our study was at risk of selection bias as BRCA mutation screening was limited by patient and family characteristics and to some extent by patients own choice of genetic counseling. Although our study urge caution when omission of adjuvant chemotherapy is considered, we are unable to estimate to what extent this could reduce mortality. We have previously, however, shown a corresponding excess mortality if adjuvant chemotherapy is omitted in patients with sporadic BC where individual trials and meta-analysis have demonstrated a benefit from chemotherapy [24,25].

Because of the rarity of germline BRCA1/2 mutations in BC [26,27], several survival studies have been conducted retrospectively and in smaller cohorts influenced by biases [28,29]. Higher quality studies in larger BRCA1/2 and population-based cohorts, with limited selection and longevity bias, have been reported more recently; Huzarski et al. [30] reported an independent and worse prognostic association of BRCA1 status with survival and a trend towards better survival after chemotherapy (HR 0.42 95% CI 0.12–1.5, p = .18) in BRCA1 carriers compared to sporadic BCs. Goodwin et al. [6] found a significantly higher risk associated with no chemotherapy treatment in BRCA2 versus sporadic BCs (HR 3.62 95% CI 1.46–8.99, p = .01) and similarly Jonasson et al. [7] reported that among patients not treated with chemotherapy, BRCA2 mutation was associated with an increased risk (HR 2.38, 95% CI 1.31–4.34, p = .005). This was not shown in the chemotherapy treated group indicating that BRCA2 BC patients derived larger benefit from chemotherapy. In summary these results are a strong indication of higher benefit from chemotherapy in BRCA BCs and for inclusion of BRCA mutation as a criteria for chemotherapy treatment. However, if these results are specific for the afore mentioned, founder mutations and age-groups, should be considered.

In conclusion, BRCA mutation type may influence risk of relapse, contralateral BC and incidence of other malignancy, but not mortality. Risk reducing contralateral mastectomy was associated with a clear survival benefit. Investigation of interaction between BRCA carrier status and effect of chemotherapy is out of the scope of this study and would require a comparable control group. However, results are in line with previous reports of the highly beneficial effect of chemotherapy in BRCA BC patients.

Acknowledgments

We would like to thank Lejla Majdanac for her excellent assistance in collecting historical data from Rigshospitalet, Copenhagen University Hospital.

Disclosure statement

AMG participated in an advisory board meeting about BRCA testing in ovarian cancers in February 2016, funded by AstraZeneca.

Additional information

Funding

The study was supported financially by Region Zealand, Region of Southern Denmark, The Danish Cancer Research Foundation, Møbelarkitekt Aksel Clausens Foundation, Landsforeningen mod Brystkraeft, Karen A. Tolstrups Foundation and Axel Muusfeldts Foundation.