0851996884.pdf

In BV-infected larvae harvested dead
under high ambient temperatures (26°C),
bacteria multiply rapidly after larval death
and contaminant bacteria may reach 2 × 108
cfu per larva. In larvae harvested alive or a
few hours after death, levels of 2 × 107 cfu
can be achieved. As a reasonable target, a
total non-pathogen microbial contaminant
level of 1 × 108 cfu ml^1 for a product with
an NPV content of 1 × 109 OB ml^1 for liquid
formulations or 5 × 108 g^1 for dry powders
would seem both attainable and defensible.
If further research confirms the essentially
benign nature of in vivo-produced BV conta-
minant microbes, then an even more relaxed
standard of 1 × 109 cfu ml^1 might be consid-
ered acceptable. Current Environmental
Protection Agency (EPA) guidelines specify
that human or non-target animal pathogens
should be absent and that if the production
system can support such pathogens then
batches must be tested for their presence.

Efficacy

Bioassays are, with microscopy, an essential
tool for quality control in BV production. In
vivoassays are central to quality control on
biological pesticides as they are the only
means by which the potency of a preparation
can be established or monitored and unac-
ceptable batches identified (Hughes and
Woods, 1986). Bioassay techniques must be
appropriate to the product in question and
several recent publications provide guidance
to standard methods both for assays and
analysis of results (e.g. Evans and Shapiro,
1997; Hunter Fujita et al., 1998; Jones, 2000).
The choice of assay depends upon many fac-
tors, but assays with neonate larvae are sim-
ple and allow many comparisons to be
carried out with relative ease (Jones, 2000).
The use of neonates may also reduce some of
the variability that may occur when older
larvae are used. In all assays, the use of labo-
ratory-reared insects from a standard culture
will produce more consistent results than
wild-collected larvae and the latter should be
avoided where possible.
However, there is an inherent, widely rec-
ognized problem: that the variability of

bioassays is very great and that the results for

a host/BV system may vary by up to three

orders of magnitude (Robertson et al., 1995).

This high variability of course creates prob-

lems if it is the intention to identify batches of

BV product whose activity is too low. To

some extent, the variability problem can be

addressed by including a standard virus

preparation of known OB concentration, and

the results recorded in terms of potency ratio

against this standard. The standard should be

a reference sample that has been character-

ized using DNA analysis. Assay variability

also improves markedly with operator expe-

rience, so that standardizing product moni-

toring should involve using specific staff who

have experience with the assay system

(Evans and Shapiro, 1997).

In seeking to set acceptable limits for

product potency based upon bioassays, it is

probably not productive to set strict a priori

limits for acceptable potency. One basis for

determining an acceptable potency standard

is for quality control limits to be developed

during the product-development cycle, as

every system will require its own criteria

(H.D. Burges, personal communication). In

the absence of such data, a guide might be

that batches of product found to have a

potency of less than 0.5 should be assayed

again. These variability problems are not

restricted to BV and fungi; similar variation

is seen in assays of B. thuringiensisand even

chemical insecticides (Robertson et al., 1995).

The most appropriate stages in the BV pro-

duction process for efficacy bioassays are on

the BV after any purification/processing but

before formulation – which is normally

called the technical product (TP) – or on the

final formulated product (FP). Bioassays on

the TP are normally preferred as the results

then allow the formulators to blend different

batches of TP, which may have different

activity, and to dilute the TP appropriately to

give an FP of specified standard activity.

Specifications

The specifications for BV products should

include the identity and amount of active

ingredient. The amount should be expressed

258 N.E. Jenkins and D. Grzywacz