Better BMT for Hurler syndrome—on the level?

“Time is brain tissue” in this autosomal recessive disease, which is generally recognized only after developmental delay, hepatosplenomegaly, dwarfism, and typical facial changes become apparent in the first 2 years of life. Untreated children become progressively neurologically devastated and generally die of heart failure, although intravenous enzyme replacement therapy can partially mitigate some non-neurologic organ complications (eg, lung, liver, skeletal). The major challenges to these children—neurologic and cardiac deterioration—can be halted, however, with swift and effective allogeneic hematopoietic cell transplantation (HCT).² 

This article and other recent studies clarify “do’s” and “don’ts” in approaching HCT for these patients. First, “do” transplant as early in life and as quickly after diagnosis as possible. Patients younger than the median age at HCT in the Boelens et al study (16.7 months) had a 16% survival advantage. Younger patients with higher intelligence measures and less organ involvement at transplant also have been shown to have better developmental and skeletal outcomes.^3,-5  With this in mind, the median delay from diagnosis to transplant of 5.2 months noted in this report should never occur. Referral to HCT should be swift, and patients should receive their procedures within weeks rather than months.

Another “do” on the list may break a long-held allogeneic HCT commandment—“transplant with the best HLA match possible, fully matched sibling donor first choice, with fully matched unrelated donor second.” Boelens et al show that event-free survival (EFS) with fully-matched siblings and 6/6 HLA-matched cords is equivalent (81% at 5 years). It can be agreed that the strongest likelihood of survival would occur with these 2 stem cell sources; however, I explain more about the risks of using a sibling later in this comment. If either of these 2 stem cell sources is not available, survival numbers fall (5-year EFS 68% for 5/6 cords, 66% for 10/10 unrelated marrow, 57% for 4/6 cords, and 41% for mismatched unrelated marrow), although survival has improved in recent years (EFS 75% for HCT after 2004). Thus, when is it worth the risk to use mismatched cords over a matched unrelated donor, and should a sibling donor always be used?

The biggest weakness noted in using noncord donors has been the high rate of partial chimerism (30%-50%). Much of this is related to earlier failed attempts at T-cell depletion and reduced-intensity conditioning approaches⁶ ; until better approaches to T-cell depletion and reduced-intensity conditioning for these patients eliminate this issue, these methods of HCT are high on the list of “don’ts.” Boelens et al further describe mixed chimerism in sibling and unrelated donor procedures in a little more than 25% of the transplants performed (see table). This outcome is assuredly better when targeted intravenous busulfan approaches are used (something the authors point to as leading to improved engraftment). However, worry remains: if the patient survives HCT and is a mixed chimera, or even a full donor chimera from a sibling donor who is a carrier, will there be sufficient levels of enzyme to deliver what the patient needs to avoid neurologic deterioration? Sibling carrier status has been linked to lower enzyme levels,⁷  and patients who do not achieve at least “normal” enzyme levels have been shown to have worse neurologic outcome in some small series.⁵  More work by this international collaborative group looking specifically at enzyme levels achieved with various stem cell sources and their effect on neurologic outcomes is promised in the article and will help clarify the importance of this measure.

The next “do” suggested by this paper is that unrelated cord blood seems to be the stem cell source of choice for most patients. If our goal is to move to HCT quickly and recipients of 5/6 cords and 10/10 unrelated donors have similar survival, the cord wins because it is readily accessible with less worry about partial chimerism using established regimens. Boelens et al show that only 7% of cord blood recipients had mixed chimerism, and 95% had normal enzyme levels (see table). This is in sharp contrast to sibling and unrelated donors, where 33% had mixed chimerism and 37% had low enzyme levels. Low enzyme expression in sibling donor recipients also occurs because many sibling donors are heterozygotes (carrier status of sibling donors was not available to the authors so correlations with enzyme level are not reported). Even when donors are not carriers (unrelated donors or cord blood), there are polymorphisms leading to marked differences in expression of enzymes associated with many inborn errors of metabolism.⁸  This fact has led some centers to choose cord blood units based on both HLA match and enzyme expression levels.

So, is a final “do” shown by Boelens et al that we should not use a matched sibling donor for patients with Hurler syndrome? Although survival for unrelated donors has improved over the past decade, this has occurred in related donor procedures as well. A choice to not use a sibling carrier may be best for long-term neurologic outcomes, but it comes with an increased risk of transplant-related mortality that a family must understand when transplant options are reviewed. More data correlating long-term enzyme expression with stem cell source and recipient outcomes is needed to bolster confidence in the existing data. How about the use of a noncarrier sibling or fully matched unrelated donor? These are certainly acceptable stem cell sources, but approaches designed to achieve full-donor chimerism must be used to ensure normal enzyme levels are achieved. Should cord blood units and sibling or unrelated donors be screened for enzyme expression or polymorphisms related to risk of lower expression? These are important study questions as research moves forward. Further study will help as we consider breaking the most sacred commandment of HCT: “sibling donors are always better than unrelated donors”—in the case of Hurler syndrome, this may not always be true.